Bitcoin 2026-09-25 daily brief — AI stance: neutral

Market-wide, open-interest-weighted funding is 0.76% annualized, effectively zero against the roughly 11%/yr exchange-default neutral baseline. The single large venue prints -1.31% annualized, about 2.07 percentage points below the aggregate, meaning shorts on that venue are marginally paying while the market as a whole sits flat. A week ago (9/16-9/19) that venue was near 0.009 percent per 8h, about 9.9%/yr, close to baseline; funding has reset from baseline to zero during a 2.9% pullback. The 14-day single-venue range is -0.0012 to 0.00947 percent per 8h (-1.3% to 10.4% annualized): never hot at any point in the rally. Funding this low is not bearish; it says there is no leveraged long to punish and no carry incentive either way.

All-venue futures OI is 55.59B USD, down 6.0B (-9.7%) from the 61.56B peak on 9/21 while price fell 2.9%: net de-leveraging. The single venue shows the same direction, 6.27B from a 6.70B peak on 9/22. Options OI is 52.2B USD, nearly equal to futures OI, after a roughly 9B step on 9/22 with 15.9B of volume that day; that step should be verified against venue coverage before being read as positioning. Skew is mild at 30 days (0.0044, 60th percentile) but put-bid at 7 days (0.014, 93rd percentile); the term structure is normal (7-to-30 slope +0.021, 30-to-90 slope +0.033 contango); 30d ATM IV is 33.9% at the 25.6th percentile of 90 days and fell 0.3 vol points over five days. Net: leverage is clean, near-term hedging demand is active, and implied vol is cheap relative to what price has been doing.

The 9/21 breakout (+6.47%) liquidated 391.6M USD of shorts, so the move to 86,597 was a squeeze. Since 9/23 the pattern flipped: longs 88.3M, 51.3M, 32.5M USD against shorts 27.6M, 25.6M, 18.3M, a long-to-total ratio of 0.56 on 9/25 with total liquidations at only the 36.7th percentile of one year. Momentum longs are bleeding out in an orderly way, not cascading. The last genuine long flush was 145.6M USD on the 9/15 -3.36% day, which set the 30d low at 75,590.

Coinbase premium -2.83 bps. The 14-day series ran from -9.46 bps (9/18) to +1.24 bps (9/22) and was negative on 12 of 14 days, with the last three at -1.22, -2.29, -2.83 bps. All of it sits well inside the +/-10 bps extreme threshold, so this is a mild, persistent offshore lead rather than regional de-risking. It is consistent with a derivatives-and-squeeze-led rally rather than US spot leadership, and it shows no change over the consolidation.

CONFLICT. The rates and dollar channel (10y 5.11%, DXY 101.04, Fed hiking) pushes against BTC; the equity and inflation-hedge channel (S&P at highs, gold 4,320, oil 105) pushes with it; on-chain and derivatives are neutral-constructive. Side taken: the local structure over the rates channel, because the measured 14-day co-movement is BTC +8.8% against a 17bp rise in the 10y and a 1.4% rise in DXY. The rates channel is real but has not been the binding constraint, and it only regains control if the 9/30 or 10/2 prints send the 10y through 5.2%.

ETF flows are the dominant micro story and they are fading. Farside data per the calendar shows a six-session inflow streak of about 2.84B USD through 9/24, but the daily BTC-denominated series decelerated 11,536 (9/21), 8,293, 4,111, 2,260, 446 (9/25): a 96% drop in four sessions, leaving the 9/25 flow at the 47th percentile, a 30d z-score of -0.46 and a flow-to-issuance ratio of 1.07, barely absorbing new supply. The 7d average of 4,870 BTC (91st percentile) is a lagging artifact of 9/21-9/22. Reading this as a fading marginal bid is a hypothesis, not validated alpha. Leverage: all-venue futures OI 55.59B USD is down 6.0B (-9.7%) from the 61.56B peak on 9/21 while price fell 2.9%, so the market is de-levering, not building; the news claim of about 2B USD added since the breakout holds only if measured from 9/17 (51.15B). Miners: Puell 0.98 (43rd percentile), hash-ribbons ratio 1.005 crossed above 1.0 on 9/20 after sitting below it through 9/19 in the supplied window, fees 640 sats per transaction at the 13th percentile. Treasury buying continues (Strive raised about 86M USD, roughly 1,001 BTC). Regulatory: the CLARITY Act cloture failed 49-50 and the 351.6M USD Bitget exploit was absorbed with altcoins rallying on 9/25. Cycle: MVRV-Z 1.00 (42nd percentile), long-term-holder SOPR 1.146 shows modest profit-taking by long holders, short-term-holder MVRV 1.153 (69th percentile) puts the implied short-term cost basis near 72,900.

hawkish. The supplied fed funds feed reads 3.63%, but the calendar records a 25bp hike to a 3.75-4.00% target range on 2026-09-16, so the feed is lagging the announced range. The 10y yield is 5.11%, up 15bp on 9/25 alone and up from 4.94% on 9/21; market pricing per the calendar implies roughly 70% odds of another hike at the 2026-10-27/28 meeting, which is outside this horizon. M2 is still growing at 5.66% year over year, so broad liquidity is not contracting even as the policy rate rises. Fear and Greed reads 71 (Greed): sentiment has not flinched at the tightening. In-window prints that feed the hike debate: PCE on 2026-09-30, nonfarm payrolls on 2026-10-02; the FOMC minutes on 2026-10-07 are outside the window.

10y at 5.11%, the high of the 14-day window, up 15bp on the day and 17bp from the 9/21 low of 4.94%. The rates spike is the freshest macro fact and the one BTC has been least tested against: it landed Friday and BTC closed -0.36%. No credit-spread feed is supplied here, so there is no credit read; this brief does not infer one.

DXY 101.04, up 1.4% from 99.61 on 9/14 and at a two-month high on hot flash PMI data and hike bets. Historically a headwind, but over the same 14 days BTC rose 8.8% (77,262 to 84,076) against the rising dollar, so the dollar channel has not been binding. The operative narrative is an inflation-hedge bid across hard assets: gold 4,320 and Brent 105.40 are rising alongside yields. A push through 101.5 combined with the 10y above 5.2% is the combination that would start to bite; at 101.0 it has not.

Risk-on. S&P 500 7,743, up 1.6% from 7,620 on 9/14, finished a winning week with the Dow up more than 470 points on Friday 9/25 despite the Treasury selloff; VIX 14.87. No Nasdaq feed is supplied here, so tech leadership cannot be split out. Equities and BTC are both ignoring the rates move, which is either a shared growth-optimism read of the hot PMI or a shared complacency; the 9/30 and 10/2 prints decide which.

Downside levels in order: 84,000 (four-session floor and the ATM strike); 81,200-81,400 (the 9/19-9/20 breakout base and the Oct 2 straddle low), the most likely test inside the window, which I put at roughly 35-40% given realized vol running above implied and long liquidations dominating; 78,800 (Oct 16 straddle low), which needs a macro-driven break, perhaps 15%; 75,300-75,600 (50MA, 30d low and the 9/15 long-flush level), a 10.4% drawdown that I put under 10% inside seven days absent an exogenous shock; 72,900 (implied short-term-holder cost basis from price divided by short-term MVRV 1.153), the level where recent buyers go underwater and short-term SOPR below 1 selling historically accelerates; 70,925 (200MA). Upside is capped near 86,600 (30d high and Oct 2 straddle high) unless ETF flows re-accelerate. These are judgment estimates, not model output.

moderate realized, low implied. DVOL is 34.62 at the 1.4th percentile of one year, and 30d ATM IV is 33.9% at the 25.6th percentile of 90 days, while realized vol is 52.2% (7d), 43.2% (30d) and 39.1% (90d), driven by the +6.47% (9/21) and +5.73% (9/18) days. Realized exceeds implied at every window, with the 7-day figure about 1.5x implied. The options market's own model-free forecast: about +/-0.7% (83,420-84,580) into the Sunday 9/27 expiry, about +/-1.75% (82,530-85,470) into 9/29, +/-3.1% (81,396-86,604) into Oct 2, which spans both PCE and payrolls, and +/-4.75% (80,010-87,990) into Oct 9. Those are distances, not directions, and given recent realized behavior they look cheap into a binary week.