Bitcoin 2026-09-20 daily brief — AI stance: bullish

Open-interest-weighted funding across venues is 0.007086% per 8 hours, about 7.76% annualized. The single major venue prints 0.006352% per 8 hours, about 6.96% annualized. Both sit below the roughly 11%/yr exchange-default neutral baseline, so long carry is neutral-to-cheap, not crowded or stretched. The 0.8 percentage-point annualized gap means that one venue runs cooler than the aggregate; whatever mild long lean exists sits on other venues. Over the 14-day window single-venue funding ranged from -0.000686% to 0.00947% per 8 hours (about -0.75% to 10.4% annualized), touching negative on the Sept 18 breakout day as shorts paid into the squeeze, then normalizing from 0.00947% on Sept 19 to 0.006352% today. Nothing here says the rally is being chased with leverage.

Market-wide futures open interest is $55.70bn, up from $51.15bn on Sept 17 and off the $56.39bn Sept 18 peak, so about $4.5bn of net positioning entered on the breakout and mostly stayed. The single tracked venue holds $6.00bn, up from $5.66bn on Sept 17, consistent with the aggregate. Options open interest is $42.9bn, the highest in the 14-day window, with roughly $14.7bn of BTC notional settling on Sept 25. Funding neutral, liquidations quiet, OI rebuilt: positioning is constructive but not one-sided, and the Sept 25 expiry is the main scheduled positioning event inside the horizon.

Two-sided flushes, now quiet. Sept 15 (-3.36%) liquidated $145.6m of longs; Sept 18 (+5.73%) liquidated $178.3m of shorts against $8.3m of longs, the largest single-side event in the window. Over 14 days longs lost about $536m and shorts about $470m, roughly balanced. Today shows $29.3m longs vs $15.6m shorts (ratio 0.53, more longs liquidated) with total liquidations at the 32nd 1-year percentile: no forced positioning is driving price at the highs.

Coinbase premium today is -1.98 bps, inside the plus or minus 10 bps neutral band. The 14-day series (in percent, so -0.0946 is -9.46 bps) was negative on 13 of 14 days: -0.34, -2.53, -0.83, -2.87, -4.20, -2.22, -2.05, -5.49, -2.19, -7.97, -8.01, -9.46, +0.13, -1.98 bps. Offshore led both the Sept 15-17 selling and the Sept 18 squeeze, with the trough on the breakout day approaching the extreme threshold. Sept 19's +0.13 bps was the first positive print in two weeks and today slipped back mildly negative. The trend is recovering from the trough but US spot has not taken leadership, which is the main reason this is not a high-confidence view.

CONFLICT. The macro tape (Sept 16 hike with more signalled, DXY 100.30, 10-year near 5%) pushes against risk assets, while the local tape (breakout to 90-day highs on rising open interest, neutral funding, 0/8 top triggers, calls bid in the front week) pushes up. Gold at 4,398 and M2 +5.41% YoY say the hard-asset bid is currently overriding nominal tightening. I side with the local tape and the hard-asset bid for this horizon, with confidence cut for the conflict.

ETF flows: Sept 18 printed +5,354 BTC (about $433m per The Block), following +2,088 on Sept 17 and outflows of -5,958 and -3,886 on Sept 15-16; the week ending Sept 18 netted only about +$6.2m after -$462.7m the prior week. The 7-day flow average was -603 BTC on Sept 18 while the 30-day average was +1,341 BTC/day. Reading that as net absorption is a hypothesis, not demonstrated research alpha. Options: about $14.7bn of BTC notional settles at the Sept 25 08:00 UTC quarterly expiry, with put/call 0.52, max pain near 72,000 and call clusters at 85,000/90,000/100,000 (crypto.news); options OI is $42.9bn, the highest in the 14-day window. Regulatory: the Clarity Act failed cloture 49-50 on Sept 15, ending Senate market-structure work for 2026; Deutsche Bank announced Bitcoin custody for EU institutions by end-2026 pending BaFin; Coinbase filed for single-stock perpetuals on Sept 18 with a 45-day CFTC window that runs past the horizon. Miners: Puell multiple 1.08 (48th percentile) is normal revenue; the hash ribbons ratio crossed above 1.0 today (1.0004) after 13 days below. On-chain activity: tx count 813,705 (99.8th percentile) with fees of 259 sats/tx (10th percentile) is high-count, low-value traffic, not fee-paying demand. Halving cycle: about 29 months after the April 2024 halving, later than the 12-18 month post-halving window where prior cycle peaks formed, but MVRV-Z 0.91 (39th percentile) and NUPL 0.34 (40th percentile) show mid-range valuation, not a cycle-top signature.

hawkish. The supplied feed shows fed funds 3.63% as of Sept 20, but the news digest reports the FOMC voted 12-0 on Sept 16 to raise the target range 25bp to 3.75-4.00%, the first hike since July 2023, with 16 of 18 participants projecting another increase this year; the feed appears to lag that decision. The 10-year sits at 4.94% after peaking at 5.01% on Sept 18 (up from 4.77% on Sept 7). M2 is still growing 5.41% YoY, so broad liquidity has not contracted despite the hike. Fear & Greed is 71 (Greed), elevated but below extreme readings. Vice Chair Jefferson speaks Sept 22 at 9:30 ET, the first Board-level remarks after the hike, and falls inside the 5-7 day horizon. The Sept 30 core PCE print and the Oct 2 payrolls report fall after the horizon and are later context, not in-window catalysts.

US 10-year yield 4.94%, up 17bp from 4.77% on Sept 7, peaking at 5.01% on Sept 18 and easing 7bp into the weekend. Direction over two weeks is higher; the last two sessions retraced. There is no credit-spread feed in this context, so no spread read is offered.

DXY 100.30, up roughly 1.5% from 98.75 on Sept 9 and holding above 100 since the Sept 16 hike. A rising dollar above 100 is the textbook headwind for BTC, yet BTC gained 38.6% off its 90-day low through this dollar advance and rose 7.4% since Sept 15 while DXY climbed from 99.68 to 100.30. The dollar is a drag on conviction, not a directional driver, at this stage.

Range-bound risk-on. S&P 500 closed 7,650.5 on Sept 18, inside a two-week band of 7,551.8 to 7,673.5, with VIX at 14.81 (down from 17.71 on Sept 16). No risk-off impulse is visible; equities absorbed the Sept 16 hike without a trend break. Sept 19-20 equity and VIX values are not supplied (weekend). No NDX feed is supplied.

First support is the Sept 22 straddle low near 80,200, then the Sept 25 straddle low near 79,300; by straddle pricing there is roughly a one-in-five chance of settling below 79,300 on Sept 25. The breakout base at 78,000-79,100 (Sept 7-9 closes) is the level that matters: losing it on a close reverses the Sept 18 gain and would be about a 1-sigma weekly move. The 30-day low at 75,590 is 6.9% below spot, about 1.65 sigma over the four trading days to expiry, roughly a one-in-twenty outcome absent a macro shock. The 50-day MA at 73,207 (-9.8%) and the short-term-holder cost basis near 71,700 (implied by MVRV-STH 1.13) sit close to the reported 72,000 max pain; reaching them inside the horizon would need a hawkish surprise from Jefferson on Sept 22 plus an expiry-driven unwind, a low-single-digit-percent probability by options pricing but the scenario that does the most damage to a long-only trend book.

low on implied, moderate on realized. DVOL is 35.65 at the 6.6th 1-year percentile and 30-day ATM IV is 34.2% (31st percentile of 90 days), while realized vol is 51.5% over 7 days, 37.0% over 30 days and 37.8% over 90 days. The options market is charging for a move of about plus or minus 1.6% by the Sept 22 08:00 UTC expiry (roughly 80,200 to 82,800), plus or minus 3.3% by the Sept 25 quarterly expiry (roughly 79,300 to 84,700) and plus or minus 5.0% by Oct 2 (roughly 78,000 to 86,000). Those are ranges, not directions. Front-week IV above 30-day IV says the next week is priced rougher than the month.

Probabilities above are derived from the supplied straddle prices under a symmetric assumption and are the options market's pricing, not a forecast of direction.