Bitcoin 2026-09-09 daily brief — AI stance: bearish

Market-wide, open-interest-weighted funding is 4.8% annualized, under the roughly 11% neutral baseline, so perp longs are paying below par and positioning is not crowded. One major venue prints 9.56% annualized, 4.75 points above the market aggregate: that venue's longs are the relatively crowded pocket, though still below neutral in absolute terms. Over 14 days the aggregate has oscillated between roughly 0 and 9.5% annualized and briefly went negative on 09-04 after the 09-03 squeeze. Funding bias is neutral.

Aggregate futures open interest is 53.2 billion, down 7.8% from the 09-03 peak of 57.7 billion and 5.8% below 08-27's 56.5 billion while price is only 4% off its high, so OI is falling faster than price: a long unwind, not fresh shorts being built. Options open interest has rebuilt to 41.5 billion from 37.6 billion on 08-31 (near the 43-45 billion prints of 08-27/28) with 3.2 billion of daily volume, so event hedging is migrating into options while perp leverage shrinks. Net: under-levered, neutrally funded, event-hedged via options. A downside move would be a spot and ETF-flow story rather than a liquidation cascade, and the upside lacks squeeze fuel because shorts were cleared on 09-03.

Today long liquidations 44.6 million against short 25.1 million (ratio 0.56, more longs), with the total at the 55th percentile of the past year, a median day. The 14-day shape is instructive: the 09-03 +5.2% day liquidated 189 million of shorts (the squeeze to 81,142), then 09-04 liquidated 80 million of longs, and 08-28's -3.2% day liquidated 103 million of longs. Each push above 80k has been followed by a long flush within one or two sessions, and longs have been the dominant casualty in 4 of the 6 sessions since 09-03. Leverage is being cleaned out of longs into a flat-to-down tape; the marginal leveraged participant has been on the wrong side.

Coinbase premium is -0.83 bps today, neutral against a ±10 bps extreme threshold. The 14-day trend is what matters: negative in 9 of 14 sessions with a mean around -0.9 bps, a trough of -5.21 bps on 09-02, and -4.22 bps on the 09-03 rally day, meaning the +5.2% move was led offshore with US spot lagging. The best positive print, +3.21 bps on 08-28, came on a down day. Read: US institutional spot is not leading at 78-81k, consistent with ETF flows turning negative. Mild and well inside the historical range, not extreme.

CONFLICT. On-chain (3 of 8 bottom indicators, 0 of 8 top, MVRV-Z 39th percentile, Reserve Risk 10th) and derivatives (OI-weighted funding 4.8% annualized, aggregate futures OI down 7.8% from the 09-03 peak, flat one-week skew) read neutral-to-constructive. The macro tape (roughly 60% hike odds, 10-year +14 bp in two weeks, S&P and VIX turning) reads risk-off. Side taken: macro, for the 5-7 day window only, because the constructive on-chain signals are 30-90 day instruments and the dated events (CPI 09-11, cloture 09-15, FOMC 09-16) all fall inside the window.

ETF flows have turned at the margin: -1,292 BTC on 09-09 (28th percentile) after -593 BTC on 09-08 (the 46.6 million dollar outflow that ended a three-day inflow run), 30-day z-score -1.02, while the 7-day average (+1,475 BTC/day, 60th percentile) and 30-day average (+1,885 BTC/day, 69th percentile) remain positive. Corporate bid persists: Strive bought 1,375 BTC at about 79,281. Structural positives: Block filed for an OCC national trust charter for bitcoin and stablecoin custody. Binary regulatory catalyst inside the window: the CLARITY Act cloture vote 09-15 at 2:15 pm ET needs 60 votes with 53 Republican seats. Sentiment drag: the Liquid Network exploit removed about 4,000 BTC (320 million dollars) and halted transactions, reportedly white-hat, so not forced supply. Miners are healthy: Puell 1.07 (47th percentile) and the hash-ribbon ratio at 0.9998, up from 0.9876 over 14 days, is one or two sessions from completing its recovery cross. Blockspace tells a two-speed story: transaction count 758,490 (99.6th percentile) with fees of 437 sats per transaction (11th percentile), so activity is batching or inscription-type volume, not economic settlement demand. Cycle: about 29 months past the April 2024 halving with MVRV-Z at the 39th percentile and 0 of 8 top indicators firing; there is no top signature in the data.

hawkish. Fed funds 3.63% (3.50-3.75% range) after five consecutive holds, and market-implied odds of a 25 bp hike at the 09-16 FOMC were roughly 60% as of 09-08. The 10-year is 4.78%, up 14 bp from 4.64% on 08-27. CPI on 09-11 carries a hot consensus of +0.4% m/m and +3.4% y/y, PPI lands 09-10, and a three-month high in crude is the cited driver. M2 is still growing +5.41% y/y, so this is a tightening response to re-accelerating inflation against an expanding money stock, not a liquidity drain. Fear and Greed at 66 (Greed) says sentiment has not priced a live hike debate.

10-year at 4.78% and rising (4.64% on 08-27, +1 bp on the day), 115 bp above the 3.63% funds rate, so the curve is steep and the term premium is building rather than the front end doing the work. There is no credit-spread feed in this snapshot, so no credit read is offered.

DXY 98.75, down from 99.70 on 08-28, about 1% softer over two weeks even as hike odds and the 10-year rose. A dollar that refuses to rally on hawkish repricing while gold prints 4,460.5 is a hard-asset bid that bitcoin has not captured (77,928 buys about 17.5 oz of gold). That is supportive on a months horizon and irrelevant on 5-7 days, where a hike surprise would snap the dollar higher and hit bitcoin with it.

Risk-off drift, not a break. S&P 500 at 7,636, down 1.4% from 7,748 on 09-03, with VIX 16.46 up from 14.32 on 09-03. Equities have started to confirm the rates repricing over the past four sessions. Triple witching on 09-18 adds a mechanical volatility node two days after the FOMC. No Nasdaq feed in the snapshot.

Levels: 77,004 (the 09-02 close, floor of the three-week range) is the first shelf; 76,442 (the straddle floor for the first post-CPI expiry on 09-12); 74,042 to 74,179 (the 09-18 and 09-25 straddle floors, about -5%); then nothing meaningful until the 50-day and 200-day cluster at 70,166 and 69,917 (about -10%). Probabilities implied by the straddles, model-free and assuming symmetry: roughly one in five for a 09-18 settle below 74,179, and roughly 5% for a close below 70,166 by 09-18 (touch odds about double). The tail is fatter than the straddle if the CLARITY cloture fails on 09-15 and the Fed hikes on 09-16, two independent negatives inside 24 hours. The cushion against an air-pocket is the clean book: OI-weighted funding at 4.8% and futures OI already down 7.8% from peak mean liquidation fuel is thin, so the expected bad path is a grind rather than a cascade. The 08-28 session (-3.2% on 103 million of long liquidations) is the template for a bad day, not the 08-19 to 08-21 spike.

moderate. Deribit 30-day implied vol (DVOL) is 40.2 at the 37.5th percentile of the past year, and 30-day ATM implied vol is 38.2% at the 77th percentile of its 90-day range, rising 1.0 point over five days. Realized vol is higher than implied at 45.2% (7-day) and 47.9% (30-day), but that is inflated by the 08-19 to 08-21 moves of +6.9%, +4.9% and +7.7%; the 90-day realized is 37.4%. The term structure is inverted at the front: 7-day IV sits 5.6 points above 30-day, so the market itself expects the coming week to be rougher than the month. Straddle-implied ranges: to Friday 09-11 08:00 UTC about ±1.65% (77,207 to 79,793), but that expiry settles before the 12:30 UTC CPI release, so it prices none of it; to 09-12 about ±2.6% (76,442 to 80,558), the first expiry carrying CPI, and the incremental variance implies roughly a ±2% move for the CPI session alone; to 09-18 about ±4.9% (74,179 to 81,821), spanning CPI, the CLARITY cloture vote and the FOMC; to 09-25 about ±6.3% (74,042 to 83,958). These say how far, never which way.