Bitcoin 2026-08-09 daily brief — AI stance: bullish

Weighted by open interest across venues, perpetual funding is running about 7.3% annualized — below the roughly 11%/yr implied by the exchange-default rate of 0.01% per eight hours. Longs are paying less than the structural baseline to hold their positions, which is a slightly under-neutral reading, not a crowded one, and it means the 3.58% advance since Aug 1 was not financed by leveraged longs. The heaviest single venue prints about 8.1% annualized, roughly 0.85 percentage points above the all-venue weighted figure; that gap says whatever crowding exists is concentrated at that one venue rather than spread across the market, but even there the level is below baseline. Two weeks ago the same weighted measure printed as high as 0.008613 per eight hours (roughly 9.4% annualized) on Jul 28 and has drifted lower since. Nothing in this positioning needs to unwind, which removes the mechanical source of the sharp downside that broke the market in early June.

Light and contracting. Market-wide futures open interest is $48.45bn, up only 1.47% from $47.75bn on Aug 1 against a 3.58% price gain — in coin terms that is a roughly 2.1% contraction, so this is a deleveraging advance. Options open interest of $25.80bn is well down from $35.99bn on Jul 31, but that is the July monthly expiry rolling off — the drop lands entirely on the Jul 31 to Aug 1 boundary — not a flight from the market, and the book has been rebuilding since ($25.26bn on Aug 1 to $27.20bn on Aug 7). The $803.7m of options volume on Aug 9 is a Sunday print and should not be read as a collapse in activity. Net: no positioning extreme in either direction, sub-baseline funding, a post-expiry options book, and no leverage overhang. Positioning is not what is constraining this market — supply is.

Quiet, and asymmetric in the bulls' favour. $12.21m of shorts liquidated against $6.64m of longs on Aug 9, a 1.84:1 ratio, with total liquidations in the 11.78th percentile of the past year. The relevant history is Jul 31, when $71.3m of longs were stopped out on the -3.0% session — the biggest single-day move in the last 30 days. That flush already cleared the leveraged longs; everything since has been a grind higher taking out shorts in small increments, with Aug 8 printing just $367k of long liquidations, the smallest in the window. There is no accumulated long-side liquidation fuel sitting under this market.

The US spot venue is trading 8.3 bps below the offshore reference, and it has been negative on every one of the last 13 sessions — from -13.8 bps on Jul 30 through -4.9 bps on Aug 8 and back to -8.3 bps on Aug 9. Against a practical extreme of ±10 bps, the level sits inside the band, so this is a persistent mild offshore lead rather than a US de-risking signal, and there is no clean trend within the window — it oscillates around -9 bps. The reading that matters is the tension with the flow data: US ETF creations ran positive five sessions straight while the US spot book still traded at a discount throughout. The consistent explanation is that creations are being sourced through OTC and block desks rather than lifting the visible order book. I read that as demand that is genuine but not price-setting — which is precisely why a positive ETF week has produced a 3.58% grind rather than a breakout.

ALIGNED — but weakly, and the weakness is the whole story. Both sides point up: macro is risk-on (S&P record, VIX 14.9, dollar down 1.68%, hike odds 55% to 42%) and the local read is constructive (4/8 bottom triggers, 0/8 top, sub-baseline funding, above the 50-day). The disagreement is not in sign, it is in magnitude — BTC converted 1.53% of a 6.03% equity move under maximum tailwind. I take the side that says the alignment is real and the conversion rate is the constraint: the macro impulse is genuine, and coin-specific supply (Strategy selling, long-term-holder SOPR at 0.859, an active wallet exploit) is eating most of it. That caps upside far more than it threatens direction, which is why I am bullish at low rather than high confidence.

This is a fight between an ETF bid and a coin-specific supply overhang, and the supply side is currently the binding constraint. On the demand side: US spot ETF creations printed positive for five consecutive sessions through Aug 7, over $750m for the week with no net-outflow days in August, absorbing 3.18x daily miner issuance on Aug 7, and the whale cohort reportedly added 19,610 BTC concentrated in the $63,000-$65,000 band — exactly where price is. On the supply side: Strategy sold roughly $105m of bitcoin last week and repurchased $81.2m of preferred, its sixth consecutive week without a purchase, so the corporate treasury bid that underwrote the prior leg has reversed into a seller. Long-term-holder SOPR of 0.859 (19.75th percentile) confirms aged coins exiting at an average 14% loss. Miner economics are stressed: Puell at 0.757 (24.7th percentile), hash ribbons at 0.9861 in the 8.86th percentile, and fees at 300.96 sats per transaction (10.09th percentile) despite transaction count at 731,687 — the 99.46th percentile. Record throughput generating near-record-low fee revenue is a miner-margin problem, not a network-health one. Two idiosyncratic overhangs are live: the Coldcard firmware exploit that has been draining wallets in waves since Jul 30, and the BIP-110 chain split at block 961,632 on Aug 8, which stalled after two blocks with miner signalling under 3% against a 55% activation threshold — contained, but a governance headline with a proof-of-work hard-fork contingency now rebased onto Knots. The CLARITY Act cloture motion filed Aug 8 positions a procedural vote after Sept 14, outside this horizon.

Hawkish, with the direction of travel softening. The effective funds rate is 3.63% and the July 28-29 meeting held the 3.50-3.75% band on a 9-3 vote — with all three dissents favouring a 25bp hike, not a cut, and Chair Warsh having stripped forward guidance from the statement. The rates market agrees: the 2-year near 4.2% sits roughly 57bp above the effective funds rate, which prices tightening rather than easing, and the 10-year at 4.69% is up from 4.61% on Jul 29 with M2 running +5.53% year over year. The softening is real but partial — the -23k July payrolls print cut market-implied hike odds to 42% from 55%. Sentiment: the Fear & Greed reading of 25 (Extreme Fear) is as of Jul 18, 23 days stale, and spans a period in which the S&P gained 6% and BTC rose. It is a lagging tint, not a live gauge, and I am not leaning on it.

The 10-year is at 4.69%, up from 4.61% on Jul 29 and at the top of its 14-day range of 4.61-4.75. With the 2-year near 4.2%, the curve is roughly +49bp and steepening on a weak payrolls print — a term-premium steepener rather than a growth steepener, which is the configuration that historically pressures the longest-duration risk assets even while equities rally. There is no credit-spread feed in this dataset, so I have no read on credit conditions and will not manufacture one; the rates signal here is unconfirmed by any credit observation.

Dollar index at 99.679, down 1.68% from 101.38 on Jul 28 and below 100 for the first time in the 14-day window. A falling dollar into a hike-odds repricing is normally the cleanest macro tailwind BTC gets. The relevant fact is how little of it BTC converted: 3.58% over the same stretch, against a 60-day drawdown that is still 19.91% from the high. The tailwind is being absorbed by coin-specific supply rather than driving price.

Risk-on, unambiguously. The S&P closed Friday Aug 7 at 7,757.64 — the last print, since Aug 8-9 are weekend — up 6.03% from 7,316.15 on Jul 29, with VIX at 14.9 on the same Friday against 20.66 on Jul 29. That is a full risk-on impulse compressed into seven sessions. BTC captured 1.53% of it over the identical window (63,874 on Jul 29 to 64,850 on Aug 7).

Implied volatility of 32.07% prices a one-sigma seven-day move of ±4.4%, or 62,140 to 67,918 — almost exactly the width of the existing 30-day range of 61,849 to 66,257. The options market is pricing range continuation, not a break, which is consistent with the low-30s implied and the contango term structure. Downside shelves, in order: the 50-day average at 63,333 (-2.6%), the 30-day low at 61,849 (-4.9%), then the 60-day low at 58,519 (-10.0%), which is roughly a two-sigma CPI reaction and the level a genuine repricing would target. Worth calibrating against recent behaviour: the largest single-day move in the last 30 days was -3.0% on Jul 31, so anything beyond about -4% would itself be a regime break rather than range noise. The long-term-holder cost basis implied by an LTH MVRV of 1.309 sits near 49,700 — far below and not in play this week. Upside is the more crowded side: resistance stacks at the 30-day high of 66,257 (just 1.85% away), then short-term-holder cost basis near 67,400 (implied by an STH MVRV of 0.965, meaning recent buyers are underwater and will sell into strength), then the 200-day at 70,141. The realistic distribution for the week is skewed toward a small grind into the 66,257 area with a fat left tail conditional on Wednesday's print.

low — at a one-year floor, which is itself the risk. Deribit's 30-day implied index is 34.93, in the 3.8th percentile of the past year, and 30-day ATM implied of 32.07% sits in the 3.3rd percentile of its rolling 90-day distribution. Realized is quieter still at the front: 9.49% annualized over 7 days against 29.17% over 30 days and 35.24% over 90. Term structure is in contango, +5.7 volatility points from 30-day to 90-day, and 25-delta skew of -0.5 points (3.3rd percentile) shows puts marginally bid — at that magnitude a tint, not a fear trade. Compression this extreme is not a stable state. With July CPI on Aug 12 inside the horizon, the base case is expansion, and the 9.49% seven-day realized number is what makes this setup fragile rather than safe.

Implied at the 3.8th percentile with a dated catalyst inside the horizon is a variance-buyer's setup, not a variance-seller's.