Bitcoin 2026-10-04 daily brief — AI stance: bullish

Open-interest-weighted funding across venues is 0.0058% per 8 hours, roughly 6.4% annualized, which sits below the 0.01%/8h (about 11%/yr) exchange-default level that traders treat as neutral. The single large venue prints 0.0065%/8h, about 7.2%/yr, roughly 0.8 percentage points annualized above the aggregate, so that venue's longs are marginally more willing to pay than the market as a whole, but neither reading is crowded, stretched or extreme. Over the last 14 days the single-venue rate ranged from -0.0019% to 0.0082% per 8h (about -2%/yr to +9%/yr) and never crossed baseline while price climbed from 83,479 to 86,490. Funding is telling us the rally has been bought mostly in spot or with modest leverage.

All-venue futures open interest is $55.3B, up from the $52.3B trough on Sep 29 but still $6.3B below the $61.6B Sep 21 peak, so leverage is being rebuilt from a flushed base; CoinDesk puts the rebuild at about $2.3B or 27,000 BTC since Sep 30, consistent with this series. The single large venue holds $6.37B, its highest since Sep 23, a venue-specific point only. Options open interest is $35.7B after the late-September expiry cut it from about $52B, and weekend options volume of $1.6B is quiet. Skew is flat at 30 days (0.002, 37th percentile) and slightly call-bid at 7 days (-0.015, 23rd percentile), with 30d ATM IV 34.2% at the 42nd percentile of 90 days and a contango term structure (90d above 30d by 2.7 vol points). Overall: constructive and uncrowded, with no evidence that longs are leveraged into the high and no put demand from hedgers.

Shorts are the forced side. Oct 4 liquidated $35.9M of shorts against $2.8M of longs (ratio 12.9), Oct 2 liquidated $100.7M of shorts against $55.0M of longs, and the Sep 21 +6.47% day wiped $391.6M of shorts. Long liquidations have been small since Oct 2 and were essentially nil on Oct 3 ($1.6M). Total liquidation volume sits at the 26th percentile of the past year, so this is a steady grind that keeps running over sellers who fade the highs, not a cascade in either direction. The pattern argues that the first move through 86,597 would be fuelled by further short covering rather than met by long liquidation.

The Coinbase premium is -1.3 bps today, inside the neutral band and well within the ±10 bps extreme threshold. The trend matters more: the 14-day series has been negative on 13 of 14 sessions, from -8.1 bps on Oct 1 to +1.2 bps on Sep 22, averaging around -2 bps. Offshore venues have led the move to the 90-day high and US spot buyers have been the slightly lagging side. This is the one flow read that leans against the bullish call: mild in magnitude, but persistent, and it is the first thing to watch for confirmation; a flip to a sustained positive premium through 86,597 would be the sign that US demand has joined the breakout.

CONFLICT. The macro tape (10y 5.24% rising, DXY 102.12 rising, a Fed that just hiked) points risk-off; the local tape (90-day high, uncrowded funding, shorts being liquidated, 0 of 8 top triggers) points up. I side with the local tape for this horizon: BTC has ignored a 30bp yield rise and a 1.7% dollar rally for two weeks, gold at $4,188 says the macro bid is in hard assets rather than pure de-risking, and the in-window macro events (ISM services Oct 5, FOMC minutes Oct 7, UMich Oct 9) are second-tier relative to the Oct 14 CPI and Oct 27-28 FOMC that sit outside the window.

Flows: spot ETF 30-day average 1,735 BTC/day and 7-day average 957 BTC/day as of the last non-null print on Oct 2 (the 7-day average has fallen from 5,035 on Sep 25); Oct 2 itself printed 2,247 BTC with a 30-day z-score of +0.14; September totalled about $2.65B per The Block. These describe current demand, and reading them as a directional signal is a hypothesis, not demonstrated alpha. Holder behaviour: long-term-holder SOPR spiked to 1.76 on Oct 4, the highest of the 14-day window, so older coins are being sold at a profit into the high; short-term-holder MVRV 1.17 and NUPL 0.15 both sit at the 71st percentile, so recent buyers are comfortably in profit and would be the first sellers on a reversal. Miners: Puell 1.09 (48th percentile), hash ribbon ratio 1.014 and rising for 14 straight days from 1.001 (26th percentile), transaction count 714k at the 99th percentile while fees are 327 sats/tx at the 10th percentile, so throughput is high and blockspace cheap. Regulatory: SEC proposed rules on Oct 2 easing fund and adviser crypto custody (supportive), while the CLARITY Act is stalled after the 49-50 cloture vote. Supply overhang: Mt. Gox still holds 34,387 BTC ahead of the Oct 31 deadline, which is later context, not an in-window catalyst. Cycle: MVRV-Z 1.07 (44th percentile) and reserve risk at the 15th percentile, mid-cycle rather than late-cycle readings.

hawkish. Fed funds 3.75% after the Sep 15-16 meeting delivered the first hike in three years to a 3.75-4.00% range; the 10y yield is 5.24%, up 30bp from 4.94% on Sep 21 and within 5bp of the Oct 2 peak of 5.29%; M2 is still growing at +5.66% YoY, so liquidity is not contracting outright. The Sep payrolls print of +29,000 with unemployment at 4.2% and downward revisions to July and August complicates a follow-up hike at the Oct 27-28 meeting (outside this horizon). The FOMC minutes on Oct 7 are the in-window read on how committed the Committee is. Sentiment gauge: Fear & Greed 65 (Greed), elevated but not extreme.

10y at 5.24%, up 30bp over 14 days and flat for the last three sessions after peaking at 5.29% on Oct 2. Direction is up, pace has stalled into the FOMC minutes and the Oct 14 CPI (which falls outside the 5-7 day horizon). There is no credit-spread feed in this snapshot, so no spread read is offered.

DXY 102.12, a 14-day high, up 1.7% from 100.39 on Sep 21 and up on 10 of the 13 available sessions. A rising dollar alongside rising nominal yields is the classic headwind for a dollar-priced asset, yet BTC gained 3.6% from the Sep 28 low over the same stretch. The dollar is therefore a drag that has not yet bitten; a push through 102.5 with yields making new highs is the setup in which it would.

Risk-on but drifting. S&P 500 last printed 7,722.72 on Oct 2, down 0.5% from 7,764.70 on Sep 21 with a 14-day range of 7,651 to 7,765; VIX last printed 15.31 on Oct 2 (14-day range 14.2 to 16.4). Oct 3-4 equity and VIX values are null (weekend). No stress signal, but equities are not confirming BTC's breakout either.

Using the straddle-implied daily sigma of about 1.76%, a one-standard-deviation move by Oct 9 is roughly 3.8%, which puts the one-sigma downside near 83,200, just under the Sep 28 low of 83,479 that serves as the invalidation level. On the options-implied distribution, a close below 83,480 inside the window is roughly a one-in-six to one-in-five outcome; that is a rough lognormal translation of market pricing, not a measured probability. A break of the 50-day MA at 79,025 (9.45% below) would be about a 2.5-sigma event within 5-7 days, a tail outcome on implied pricing but the kind that vol compression at the 14th percentile has produced before. The first supports are the Oct 1-3 and Sep 23-27 congestion between 84,076 and 84,842, then 83,480, then the 50-day MA at 79,025 and the 30-day low at 75,590 (-12.6%). The downside path most consistent with the contradicting signals would be a hawkish FOMC minutes reaction on Oct 7 that lifts the 10y above 5.29% and DXY above 102.5, with long-term-holder distribution (SOPR 1.76) meeting it.

low. DVOL 36.5 sits at the 13.7th percentile of the past year; realized vol is 21.0% over 7 days, 37.3% over 30 days and 38.3% over 90 days, so the one-week realized print has collapsed to roughly 60% of 30-day implied (34.2%), with implied barely moving over five days (+0.05 vol points). The at-the-money straddle market charges ±1.55% for the Oct 6 expiry (breakevens 85,159 to 87,841), ±2.05% for Oct 7 (84,727 to 88,273) and ±3.0% for Oct 9 (84,384 to 89,616); ±4.8% by Oct 16 (82,827 to 91,173). That is the options market's forecast of how far price travels, and it says nothing about direction. Vol this compressed at a price high is a setup for expansion, which is why the invalidation level below is kept tight.

Levels are from supplied price history and moving averages; probabilities are translations of the supplied options-implied move and carry no research certification.