Bitcoin 2026-08-06 daily brief — AI stance: bearish

Perpetual longs are paying almost nothing. The open-interest-weighted rate across venues annualizes to 2.74%, roughly a quarter of the ~11% that the exchange-default rate represents as a neutral baseline. That is not crowded, not stretched, and arguably under-positioned. A single major venue prints 3.77% annualized, about 1.0 percentage point above the cross-venue weighted figure, so what modest long-side lean exists sits on that one venue rather than where the bulk of open interest resides — and even there it remains well under the neutral mark. The practical consequence cuts both ways: there is no leveraged long overhang whose forced unwind would produce a violent drop, but equally there is no positioning imbalance left to squeeze on the upside beyond the covering that already happened.

Market-wide futures open interest is $48.57B, essentially flat against $48.75B on July 25 and down from $49.50B on Aug 5. The 2.42% recovery off the Aug 1 low therefore came with flat-to-lower open interest — covering and spot, not new leverage. Structurally healthier than a leveraged bounce, but it also means no imbalance remains to squeeze. The single-venue open interest figure of $1.44B is one exchange, roughly 34 times smaller, and tracked the same flat pattern. Options open interest of $27.12B is rebuilding from the $25.26B post-expiry trough on Aug 1 after the July monthly rolled $35.99B off the board on July 31 — a normal reset, not a positioning signal. The standout is implied volatility: 30-day ATM at 32.93% sits in the 6.7th percentile of its rolling 90 days and DVOL at 34.98 in the 4.1st percentile of the year, with 25-delta skew of 0.026 technically put-over-call but at the 3.3rd percentile of its own 30-day range. Puts still carry a premium in absolute terms, but hedging demand has drained to the bottom of its recent range. Nobody is paying for protection into a payroll print and a CPI print.

Total liquidations on Aug 6 were $16.9M — $8.18M long against $8.69M short — sitting at the 10.7th percentile of the past year. Very quiet. The two-week pattern is the informative part: two large long-liquidation days, $71.7M on July 27 and $71.3M on July 31, marked the flush into the Aug 1 low of 62,780. The flow then reversed, with $29.0M of short liquidations on Aug 4 and $37.9M on Aug 5 carrying price back to 64,580. So the recovery leg was short-covering, not fresh long demand. Both sides have now spent their fuel, which is consistent with the volatility compression.

The Coinbase premium at -9.62 bps sits essentially at the ±10 bps practical extreme, but the snapshot matters less than the streak: all 13 sessions with data printed negative, ranging -8.42 to -13.78 bps and averaging -10.03 bps, without a single positive tick. Offshore has led the US bid for two straight weeks. That directly conflicts with the ETF inflow narrative — spot ETFs took $211.5M on Aug 4 and the 7-day flow average is +1,397 BTC, yet US-hours spot trades at a persistent discount. The resolution is mechanical: ETF creations clear through authorized participants and OTC desks and do not have to lift the visible US order book, while the broader US spot tape is a net seller. Read it as the ETF bid being real but narrow — IBIT alone was $170.3M of the $211.5M on Aug 4 — a floor under price rather than a driver of it.

CONFLICT. The on-chain and derivatives read is constructive — BOTTOM ZONE at 4/8 bottom and 0/8 top, Reserve Risk at the 2.5th percentile, funding well below neutral, more shorts than longs liquidated, price above the 50-day. The macro read is a Fed pricing a coin-flip September hike into a CPI consensus that re-accelerates, combined with demonstrated failure to convert dollar weakness and an equity melt-up into price. I side with macro.

Three threads. Flows: US spot ETFs turned genuinely positive for four sessions — roughly $170M on Aug 3, $211.5M on Aug 4 (of which IBIT alone was $170.3M), about $243M on Aug 5 — but Aug 6 collapsed to roughly $9M. The 7-day flow average of +1,397 BTC sits at the 57th percentile while the 30-day average is still net negative at -491.7 BTC/day, in the 16th percentile. Four good days have not turned the month. Miner economics: hash ribbons at 0.987 in the 9.3rd percentile means 30-day hash is running below 60-day — capitulation is in progress, not finished. Puell at 0.670 (18.1st percentile) and fees at 675.6 sats/tx (12.7th percentile) despite transaction count at the 95.4th percentile means high block-space throughput with no fee competition, so no revenue relief. MARA's Q2 confirmed the pressure: a $611M net loss against consensus for a $0.35 profit, revenue down 27% y/y, and treasury down 29% y/y to 35,577 BTC. Public miners are net sellers. Regulation: the CLARITY Act's Senate floor window closes with today's session — no cloture motion filed as of Aug 6, absent from the floor calendar, recess starting Aug 10, and 60 votes required. The next realistic window is September or a lame-duck session. Offsetting all of this, whale wallets excluding exchanges and mining pools have grown to roughly 3.06M BTC from about 2.87M in December 2025, with accumulation accelerating below $60,000.

Hawkish, but the market is genuinely split and today resolves it. The funds rate sits at 3.63% after the July 28-29 meeting held the 3.50-3.75% range, and fed funds futures as of Aug 4 priced a 56.9% probability of a September HIKE against 43.1% for a pause, with Kalshi at 49% hold versus 47% for 25bp. Against that, the 10-year rallied from 4.75% on Aug 3 to 4.63% on Aug 6 after ADP printed 44k private jobs versus the ~70k forecast — rates are trading the soft-labor case while futures price the tightening case. Today's payroll print (consensus +83k) breaks the tie. The July CPI consensus for Aug 12 compounds the hawkish side: +2.8% y/y headline and +3.0% core against June's 2.7% and 2.9% is re-acceleration on both measures. M2 growth of 5.53% y/y is the one genuinely supportive input. On sentiment, the Fear & Greed reading of 25 (Extreme Fear) is dated July 18 — three weeks stale, describing the market before the S&P's 4% advance and the VIX collapse from 20.66 to 15.15. Treat it as a lagging input, not a live gauge.

The 10-year at 4.63% is down 8bp from 4.71% on July 25 and 12bp from 4.75% on Aug 3, a rally driven by the weak ADP print. With funds at 3.63%, the funds-to-10s curve is roughly 100bp positively sloped. There is no credit-spread feed in this dataset, so no credit read is available — I am not going to infer one from equity vol.

DXY at 99.954, having fallen from 101.502 on July 27 to 99.666 on Aug 5 before ticking back up. The break below 100 is a meaningful 1.53% move and is normally a first-order tailwind for BTC. It did not act like one — over that same window BTC gained 0.87%. A dollar decline of that size that produces sub-1% appreciation is a tailwind the asset is declining to use.

Risk-on and unambiguous. The S&P 500 at 7,709.96 is up 4.00% from 7,413.18 on July 27, near the top of its recent range, with the VIX at 15.15 down from 20.66 on July 29. Oil fell on the Aug 3 US-Iran Hormuz arrangement, removing a geopolitical risk premium. Note the feed carries the S&P only — there is no Nasdaq series here, so this is a broad-market read, not a tech-specific one.

Not a crash setup — a range-resolution setup, and the range lows are close. With 30-day realized volatility at 30.68% annualized (about 1.61% per day), a one-sigma week from 64,303 lands near 61,570, which is effectively the 30-day low of 61,849. A two-sigma week lands near 58,840, essentially the 60-day low of 58,519. So the entire visible downside structure sits within two standard deviations, which is why I am calling direction rather than magnitude. First support is the Aug 1 and July 31 lows at 62,780 and 62,888, immediately beneath the 50-day at 63,244 — a cluster tight enough that breaking one likely breaks all three. Below 61,849 the next real reference is 58,519, and below that there is nothing structural on the recent tape. Two qualifiers that cap the tail: funding at 2.74% annualized means no leveraged long stack to cascade, and Reserve Risk at the 2.5th percentile with whales at roughly 3.06M BTC means long-horizon buyers are present under the market. The realistic bad week is 61,500-62,800, not a capitulation leg.

low