Bitcoin 2026-09-17 daily brief — AI stance: bearish

Open-interest-weighted funding across exchanges is 7.45%/yr, below the roughly 11%/yr exchange-default baseline, so the market as a whole is neutral to slightly cool on the long side. The single major venue prints 9.03%/yr, 1.59 percentage points above the aggregate, so what modest long lean exists is concentrated on that venue rather than market-wide. Funding drifted up from about 1-3%/yr over 2026-09-04 to 2026-09-07 to a 5-10%/yr range from 2026-09-08 onward, so longs re-engaged modestly after the +5.01% day on 2026-09-03 but never reached crowded territory.

Market-wide futures open interest is 51.15B USD, down 6.8% from 54.91B USD on 2026-09-04, so leverage was reduced into the drawdown rather than added. The single major venue holds 5.66B USD, roughly flat over 14 days. Options OI is 40.9B USD with 14.63B USD of BTC notional expiring at 08:00 UTC on 2026-09-25 at a put/call OI ratio of 0.52 (call-heavy) and max pain at 72,000-75,000. Skew shows modest downside hedging: 30d 25-delta put-over-call at +1.03 vol points (77th percentile of 30d), 7d at +0.55 (83rd percentile). The 7-30 day term slope is +1.76 vol points, so no near-term event premium is priced despite the expiry. Net: lightly positioned, mildly hedged, with a call-heavy quarterly expiry whose max pain sits below spot.

The -3.36% session on 2026-09-15 forced 145.6M USD of long liquidations against 27.9M USD of shorts, the largest long flush of the 14-day window. On 2026-09-17 the pattern flipped: 20.0M USD of shorts versus 9.0M USD of longs (ratio 2.22) on the post-Fed bounce. Total liquidations sit at the 18th percentile of 1y. Read: two-sided chop with leveraged longs already flushed once this week and no cascade in progress.

Coinbase premium is -8.01 bps, the most negative print of the 14-day window and just inside the -10 bps extreme-decile threshold. The trend is the point: it moved from +1.8 bps on 2026-09-05 to -0.8 bps on 2026-09-09, -4.2 bps on 2026-09-11, -5.5 bps on 2026-09-14 and -8.0 bps on both 2026-09-16 and 2026-09-17. Offshore spot is leading and US spot is not absorbing, which lines up with the three-day ETF outflow of about 10,161 BTC.

CONFLICT. The macro tape (FOMC hike on 2026-09-16, 10y 5.00%, DXY 100.22) pushes bearish. The local trend structure (above both MAs, neutral 7.45%/yr funding, all-venue OI down 6.8% to 51.15B USD, cycle monitor BOTTOM ZONE) is neutral to constructive. The local flow and holder-behavior data (7d ETF average at the 10.6th percentile, Coinbase premium -8.01 bps, SOPR below 1) side with macro. I side with macro and flows for the 2026-09-18 to 2026-09-24 window: a 5% 10y and a hiking Fed dominate a 5-7 day horizon, and the trend gate is a lagging structure sitting 5.85% below spot.

US spot ETF flow is the live micro story: -316.9 BTC on 2026-09-17 after -5,958.4 and -3,885.9 on the two prior sessions, roughly -10,161 BTC over three days. The 7d average is -1,930.7 BTC at the 10.6th percentile, the 30d average is still +1,153.6 BTC (50.7th percentile), and Farside reported -462.7M USD for the week ended 2026-09-11, ending a three-week inflow streak while September remains about +307M USD. Whether these outflows lead price is a hypothesis, not demonstrated alpha. On-chain, SOPR has printed below 1 for three straight days (0.9973, 0.9993, 0.9985), aSOPR is 0.9968, and short-term-holder NUPL fell to 0.066 from 0.118 on 2026-09-06, so recent buyers are at breakeven. Miner economics are soft but not distressed: Puell 0.94, hash ribbons 0.9994 (15th percentile) with 30d hashrate marginally below 60d, fees at 430 sats/tx (10.7th percentile) despite tx count at 652,518 (98.2nd percentile). Regulatory: the Clarity Act cloture vote failed 49-50 on 2026-09-15, and the House Financial Services Committee advanced the Strategic Bitcoin Reserve bill H.R. 8957 by 28-21 on 2026-09-16 with no floor vote scheduled and no new purchases. The Deribit Q3 expiry at 08:00 UTC on 2026-09-25 carries 14.63B USD of BTC notional, put/call OI 0.52 and max pain at 72,000-75,000, below spot. Dated catalysts inside the 2026-09-18 to 2026-09-24 window are thin: daily Farside ETF prints starting 2026-09-18 and weekly jobless claims on 2026-09-24. The quarterly expiry falls the morning after the window closes, and PCE on 2026-09-30 and payrolls on 2026-10-02 are later context, not in-window catalysts.

hawkish. The feed shows fed funds at 3.63% and the calendar records a unanimous 25bp FOMC hike on 2026-09-16 to a 3.75-4.00% target range, the first since 2023, with a dot-plot median of 4.1% for end-2026 and end-2027 and 16 of 18 participants projecting further hikes this year. The 10y sits at 5.00%, up 23bp from 4.77% on 2026-09-05, and the 2y printed 4.744%, its highest since July 2024. M2 is still growing 5.41% YoY, so this is a rate-tightening regime rather than a money-supply contraction. Fear & Greed is 50 (Neutral) and VIX is 15.44, so sentiment has not yet priced the tightening.

10y yield 5.00%, the highest of the 14-day window and up 23bp since 2026-09-05. The news notes it snapped an eight-day rising streak intraday on 2026-09-17 but the feed still closed at 5.00. Real-rate pressure is the transmission channel to watch. There is no credit-spread feed supplied, so no credit read is possible here.

DXY 100.22, up about 1.5 points from 98.75 on 2026-09-09 and holding above 100 for a second session (100.32 on 2026-09-16). A rising dollar alongside a 5.00% 10y is the standard headwind for a non-yielding asset. Gold at 4,399.7 while BTC sits 6.02% below its 30d high says the hard-asset bid in this window is going to gold, not BTC.

Risk-on-lite relief inside a two-week drift. S&P 500 at 7,637.76 is 1.05% below the 7,718.60 print on 2026-09-04 but rose 1.1% on 2026-09-17 after the Fed decision, with a chipmaker gauge up 3% and VIX falling from 17.71 to 15.44. BTC fell 4.1% over the same 2026-09-04 to 2026-09-17 span, so BTC is underperforming equities rather than tracking them.

Base case for 2026-09-18 to 2026-09-24 is a retest of the 75,590 close from 2026-09-15 (about -1.0% from 76,371), with the straddle-implied one-sigma floor for the 2026-09-25 quarterly at 74,414 (-2.6%). A slide to the 50MA at 72,148 (-5.5%) would land at the top of the 72,000-75,000 max-pain band and is roughly a 1.6-sigma move on the 2026-09-25 straddle, plausible but not the base case. A break below 72,000 toward the 30d low at 69,291 (-9.3%) would require the vol regime to shift off its 1y low and most likely an equity risk-off from the current VIX 15.44, which is not the expected path in this window. Upside tail is 78,000-78,200 then 80,329-81,265.

low. DVOL is 33.59 at the 0th percentile of its 1y range, 30d ATM IV is 32.8% at the 14th percentile of 90d and fell 2.4 vol points over five days. Realized vol is 31.2% over 7d, 49.0% over 30d (inflated by the +6.88%, +5.24% and +7.00% sessions of 2026-08-19 to 2026-08-21) and 36.4% over 90d. The Deribit straddle market, with the index at 77,190 as of 2026-09-18 04:16 UTC, charges about +/-1.3% (75,977 to 78,023) for the 2026-09-19 expiry, about +/-3.35% (74,414 to 79,586) for the 2026-09-25 quarterly, and about +/-4.9% (73,237 to 80,763) for 2026-10-02. Those ranges say how far the market expects price to travel, not which way. Implied vol at a 1y low into a 14.6B USD expiry and a hiking Fed is a compression that more often resolves by expanding than by compressing further.