Bitcoin 2026-09-23 daily brief — AI stance: neutral

Open-interest-weighted funding across exchanges runs 2.8% annualized, well under the roughly 11% per year that corresponds to the exchange default and marks neutral. Longs are not paying meaningfully to stay long. One major venue prints 8.53% annualized, 5.7 percentage points above the all-venue aggregate, so what leverage-long appetite exists sits concentrated on that venue, but even there the reading is under the neutral baseline and cannot be called crowded or stretched. Over the fortnight the aggregate ranged 0.0026% to 0.0094% per 8 hours, all within the ordinary band. Funding is therefore not the mechanism that ends this move; the mechanism is spot distribution and hedging, not carry.

All-venue futures open interest is $57.8bn, down 6.1% from the $61.6bn peak on 2026-09-21 but still 8.7% above the $53.2bn of 2026-09-10, so part of the squeeze leverage has been flushed and part remains. Options open interest is $52.9bn after a $9bn jump on 2026-09-22 on $15.9bn of volume, positioning and rolls ahead of the 2026-09-25 08:00 UTC quarterly expiry that removes roughly $16bn, about 37% of the largest venue's bitcoin book, followed by CME settlement at 15:00 UTC the same day. The expiring book is call-heavy, put/call 0.69 with max pain at 75,000 far below spot, while one-week put skew at the 97th 30-day percentile shows hedges being bought into it. Net: not crowded on funding, but leveraged, churning and hedged, the profile of digestion rather than breakout continuation.

The rally was a short squeeze and the pullback is a long flush. Shorts lost $178.3m on 2026-09-18, the +5.73% day, and $391.6m on 2026-09-21, the +6.47% day. Since then longs have been the casualty: $55.9m on 2026-09-21, $28.4m on 2026-09-22 and $88.3m on 2026-09-23 against $27.6m of shorts, a 0.31 ratio. Total liquidations sit at the 78th one-year percentile, so churn is elevated. The earlier -3.36% day on 2026-09-15 cleared $145.6m of longs, meaning the leveraged long base was reset twice in nine days; that limits cascade fuel below, consistent with a retest of the breakout base rather than a collapse through it.

Coinbase premium is -1.22 bps today, neutral on its own. The fortnight trend is the story: the spread was negative on 12 of 14 days, deepening from -2.87 bps on 2026-09-10 to -7.97, -8.01 and -9.46 bps on 2026-09-16 through 2026-09-18, just short of the -10 bps extreme, then snapping to +0.13 bps on 2026-09-19 and hovering near flat since, +1.24 bps on 2026-09-22 and -1.22 bps today. Offshore led the selling into the 2026-09-15 low, and offshore or derivatives, not US spot, led the rally: even on 2026-09-21, the record ETF creation day, Coinbase traded at a discount. That fits ETF creations executed through authorized participants and OTC desks rather than Coinbase spot, and it means the US bid that historically accompanies durable institutional legs is not visible in this spread. A move above +5 bps would change that read.

CONFLICT. The rates and dollar tape, a Fed that hiked on 2026-09-16 with more signalled, a 10-year at 4.96% and a DXY at 101.09 rising, pushes against risk assets. The on-chain and derivatives tape, price 13.2% above the 50-day, OI-weighted funding at 2.8% annualized, cycle neutral, ETF 7-day flows at the 75th percentile, points up. Equities at record and gold at $4,323 show the market currently treating bitcoin as a hard asset rather than rate-sensitive tech. I side with the macro headwind for this window's tilt, because the squeeze fuel that carried the 14.6% move is spent, but with the local trend for the regime: pullbacks inside the window are digestion, not reversal, unless 78,173 breaks.

Flows: spot ETF creations printed 11,536 BTC on 2026-09-21, about $999m and the largest session of 2026, then 8,293 on 2026-09-22 and 2,141 on 2026-09-23; the 7-day average is 2,795 BTC per day at the 75th percentile and the 30-day 2,092 at the 74th, with today's flow 5.3 times daily issuance. The same series printed -5,958 on 2026-09-15 and -3,886 on 2026-09-16, so flows are chasing price in both directions; any directional read of that is a hypothesis, not validated alpha. Strategy resumed purchases with about $75m in the week of 2026-09-14. Holders: long-term-holder SOPR at 1.4586 on both 2026-09-22 and 2026-09-23 against 0.78 to 1.20 earlier in the fortnight says long-term holders are realizing profit into strength; short-term-holder SOPR slipped to 0.99957, the first sub-1 print since 2026-09-15, and the 247wallst piece notes aggregate ETF holders reached break-even on 2026-09-21, so the marginal buyer of the last week sits near cost. Miners: Puell 0.95 at the 40th percentile, fees at the 11th percentile, hash ribbons crossed back above 1.0 on 2026-09-20 and now read 1.0033, an easing of miner stress. Regulatory: the CLARITY Act cloture vote failed 49-50 on 2026-09-15, stalling the market-structure catalyst. Cycle: 29 months past the April 2024 halving, with the cycle monitor neutral and no top indicator firing.

hawkish. The supplied effective fed funds reading is 3.63% against a 10-year at 4.96% and M2 growth of 5.66% year on year. The news calendar records the FOMC's unanimous 25bp hike to 3.75% to 4.00% on 2026-09-16, the first since 2023, with 16 of 18 participants projecting at least one further 2026 increase, so the 3.63% print appears to predate the hike. Chair Warsh's language is being parsed for how much further tightening is intended, and Brent near $99 keeps the inflation argument alive. Fear & Greed at 71, Greed, says crypto sentiment has not priced that. FOMC minutes arrive 2026-10-07, after this window.

10-year at 4.96%, up from 4.80% on 2026-09-10 with two touches of 5.01% on 2026-09-18 and 2026-09-22; the 2026-09-22 market wrap put the 2-year at 4.76%, so the curve is positively sloped with policy at 3.75% to 4.00%. No credit-spread feed is supplied, so no spread read is offered.

DXY 101.09, up 2.0% from 99.10 on 2026-09-10 and up four straight sessions from 100.21 on 2026-09-18. Bitcoin rallied 10% through that climb, so the dollar has not been the marginal driver of the move, but a strengthening dollar paired with a 5% 10-year is the classic headwind and removes the liquidity tailwind that usually follows breakouts. A push through 102 inside the window would matter more than today's level.

Risk-on. The S&P 500 at 7,706 is 0.8% below the 2026-09-21 record close of 7,764.70, a day it rose 1.5% with the Nasdaq up 2.3%; VIX 15.18 versus 17.8 on 2026-09-10. Equities are absorbing a hiking Fed because oil retreated to $99 and yields slipped from 5.01%. The soft underbelly is the real economy: UMich sentiment 47.8 preliminary with the final due 2026-09-25, Q2 GDP 1.5% annualized, ADP private payrolls +38k in August. A data-driven equity wobble on 2026-09-30 is the fastest channel from macro to bitcoin.

By construction a bit over 40% of outcomes fall outside the ±4.0% straddle band of 80,629 to 87,371 by 2026-10-02, split about evenly; my tilt puts more of that mass below. Layers down from 84,382: 83,089, the one-day straddle low; 81,236 and 80,874, the 2026-09-19 and 2026-09-18 closes that form the breakout base; 80,629, the 2026-10-02 straddle low; 79,344, the 2026-10-09 straddle low; 78,173, the 2026-09-14 close; 75,590, the 2026-09-15 close and 30-day low; 74,530, the 50-day. A retest of 80,600 to 81,200 is my base case inside the window, roughly one in two versus about two in five implied by the options. A close below 78,173 is a minority case, perhaps one in five, and would need a macro shock such as a hot 2026-09-30 PCE or a summit failure. The 50-day at 74,530, 11.7% below, lies outside the implied band for the window and would require the cascade that the twice-reset long base makes less likely.

moderate, with realized running hot above implied. DVOL 35.81 sits at the 6.3rd percentile of the last year and 30-day ATM implied vol is 36.3%, the 60th percentile of the last 90 days and up 2.9 vol points in five days, yet 7-day realized is 62.7% and 30-day realized 43.3%. The straddle market charges about ±1.7% by the 2026-09-25 08:00 UTC expiry, a band of 83,089 to 85,911; about ±4.0% by 2026-10-02, 80,629 to 87,371; and about ±5.5% by 2026-10-09, 79,344 to 88,656. The one-day implied daily sigma of 1.9% compares with a 7-day realized daily sigma near 3.3%, so options are pricing a calmer tape than the one just delivered. The 7-to-30-day term slope is positive at 1.4 vol points, the ordinary shape, so no event premium is loaded for the summit or the expiry. These figures say how far price may travel, not which way.