Bitcoin 2026-08-05 daily brief — AI stance: neutral

Subdued, not crowded — and this is where a careless read would go wrong in both directions. Open-interest-weighted funding across venues annualizes to 3.72%, from 0.0034% per 8 hours. The exchange-default 0.01% per 8 hours that defines neutral annualizes to roughly 11%, so the market as a whole is paying about a third of baseline to hold long exposure. The single major venue in this dataset reads 3.49% annualized, 0.23 percentage points under the all-venue aggregate — a trivial gap, and with no crowding present there is nothing to locate. The trajectory matters more than the level: the same all-venue measure ran 0.0086% per 8 hours on Jul 28-30 and has more than halved into today. Nobody is paying up for leverage. That removes the long-liquidation fuel that has capped every bounce inside this range, and it equally removes the leveraged-short fuel that would be needed to squeeze price through 66,257. Cool funding here is a reason to expect a smaller move, not a bigger one.

Light and rebuilding, with no overhang on either side. Market-wide futures open interest is $49.50bn, up from $47.22bn on Jul 28 and the highest of the last two weeks — positions are being re-established near the top of the range, but with funding at 3.72% annualized that rebuild is not directional crowding. (The single-venue open-interest series in this dataset reads $1.45bn and runs roughly 30 times smaller; it describes one venue, not the market.) Options open interest is $26.51bn, recovering from $25.26bn on Aug 1 after the Jul 31 month-end expiry rolled roughly $10.7bn off the $35.99bn peak — so $26.5bn is a rebuilding book rather than a shrinking one. Worth noting that at $26.5bn against $49.5bn of futures open interest, this dataset does not corroborate the widely-reported claim that BTC options open interest has overtaken futures; on these numbers it is about 54% of it. Term structure is in contango with 90-day implied vol above 30-day (slope +0.0588) and 25-delta skew at 0.0453, the 43rd rolling percentile — balanced, with no crash bid being paid for. Net: flat-to-light positioning, cheap volatility, no leverage to liquidate.

Asymmetric toward shorts and small in absolute terms. $37.9m of short liquidations against $6.6m of long liquidations today is a 5.75:1 ratio, and it is the third consecutive session in which shorts wore the damage — Aug 4 was $29.0m against $6.1m. But total liquidation activity sits at only the 32nd percentile of the last year, so this is attrition rather than a cascade. The context that matters is the other side of the range: on Jul 27 and Jul 31 longs took $71.7m and $71.3m respectively. Leverage has now been flushed in both directions inside two weeks, which is precisely why 61,849-66,257 is holding. There is no trapped cohort left to force a resolution.

This is the most persistently bearish item in the entire snapshot and the reason I will not upgrade the view. The Coinbase premium is -8.74 bps today and has been negative on every one of the 13 reported sessions in the last 14 days, averaging roughly -9.9 bps and reaching -13.78 bps on Jul 30. A two-week average sitting essentially on the ±10 bps extreme-decile threshold is not noise — offshore venues have led price on every single session of this range. What makes it decisive is that the ETF inflows did not move it: Aug 4 printed +3,293 BTC of creations and the premium still read -8.42 bps. Creations without a US spot premium reads as authorized-participant and basis mechanics rather than a broad US institutional bid stepping in. That is the strongest available argument against treating the 50-day reclaim as the start of something, and it is also the cleanest thing to watch: if the premium turns positive, the thesis in this brief is wrong.

CONFLICT

Three crypto-specific items, and together they explain the macro disconnect. First, ETF flows have turned but not repaired: +2,677 BTC on Aug 3, +3,293 on Aug 4 and +737 on Aug 5 lifted the 7-day average to +829 BTC, yet the 30-day average is still -855 BTC per day at the 12.7th percentile, and Farside recorded $61.5m of net weekly outflows for the week ended Jul 31. Today's flow absorbed 1.66 times miner issuance, 293 BTC of excess absorption — positive, but one day against a month of distribution. Second, the CLARITY Act market-structure bill has no confirmed floor vote and Aug 7 is the Senate's last scheduled workday before the Aug 10 recess. Cloture needs 60 votes, roughly seven Democrats on a united Republican caucus, and two Democrats have conditioned support on an ethics provision barring senior officials from crypto business ties; as of Aug 4 no cloture motion was confirmed filed. That is a binary, unresolved, crypto-only catalyst inside the horizon, and it is the kind of thing that keeps a US institutional bid on the sidelines. Third, miner economics are stressed: the Puell Multiple is 0.745 at the 23.6th percentile and the hash-ribbons ratio is 0.987 at the 9.4th percentile — below 1, meaning short-term hashrate has rolled under its long-term trend — with MARA reporting Q2 on Aug 6. Separately, the Coldcard firmware exploit removed 1,596 to roughly 2,055 BTC across more than 7,300 addresses via a weak random-number generator in a March 2021 build. That is a self-custody confidence hit and a narrative negative, not a supply event of consequence at this size.

Neutral, with the long end doing the tightening the Fed is not. The policy rate is 3.63% and the July 28-29 meeting was a hold, with minutes not released until Aug 19. Against that, the 10-year is 4.63% after printing 4.74% last week — its highest since January 2025 — leaving roughly 100bp of positive spread over funds. M2 is expanding at +5.53% year-over-year, so the quantity of money is growing while the price of duration rises; that combination is ambiguous for a long-duration non-yielding asset, not supportive. Nothing on the calendar resolves stance before the Aug 12 CPI. The nearer input is the Aug 7 July payroll report, consensus around +80k against +57k in June with the unemployment rate possibly rising to 4.3% from 4.2%, and with revisions to the prior two months a stated focus. The live sentiment gauge reads 25, Extreme Fear, but that print is dated 2026-07-18 and is therefore 18 days stale — I weight it as context, not as a current reading.

The 10-year is 4.63%, easing from 4.75% on Aug 3 and 4.70% on Aug 4, but the week's 4.74% high was the highest since January 2025, so the two-week direction at the long end is sideways-to-higher even as the short end holds at 3.63%. That is a term-premium story rather than a policy story, and high long real yields are historically the specific macro variable that caps a long-duration asset like BTC even while equities rally on earnings. There is no credit-spread feed in this dataset, so I have no read on credit conditions at all — the rates view rests on the level and the slope alone and should not be extended to risk appetite in credit.

DXY is 99.666, down from 101.465 on Jul 24 — a 1.8% two-week slide that broke the 100 handle. At this point in the cycle that is textbook BTC-supportive. Over the identical window BTC went from 64,113 to 64,580, a gain of 0.73%. The dollar signal is real; the transmission into BTC is not, and that gap is the single most important observation in this brief.

Risk-on, unambiguously. The S&P 500 is 7,723.55 after a record 7,736.52 close on Aug 4, a session in which it gained 1.8% and the Nasdaq 2.6% on AI-linked large-cap results and two consecutive days of falling crude as the US, Iran and Oman negotiate Strait of Hormuz transit arrangements (Brent settled below $80, WTI near $75). VIX is 15.81, down from 20.66 on Jul 29. Equities are making all-time highs while BTC sits 21.5% below its own 90-day high — the correlation is not merely weak here, the sign is wrong.

Three tiers, and collapsing them into one percentage would misrepresent all of them. The near tier is the 50-day at 63,244 (-2.1%) and the Aug 1 low at 62,780 (-2.8%). A routine event-day move reaches both, and with total liquidations at only the 32nd 1-year percentile and funding at 3.72% annualized there is no forced-selling overhang to accelerate through them — these are gate-flip levels far more than they are damage levels. The middle tier is the 30-day low at 61,849 (-4.2%). Losing that means the two-week range has failed, and because the Coinbase premium has now been negative for 14 consecutive sessions there is no visible US spot bid between there and the 60-day low at 58,519 (-9.4%) — the gap is real and it is the specific reason a range break would travel further than the break itself warrants. The far tier requires two things going wrong together: a CLARITY cloture failure or slip past recess combined with an Aug 7 payroll print that pushes the long end back toward last week's 4.74%. That pairing is what puts 58,519 in play inside a 5-7 day window; with 30-day realized vol at 30.88% a -9.4% weekly move is roughly a 2-sigma event, so unlikely rather than a tail. Against all of it, the downside is cushioned by valuation rather than by flow — Reserve Risk at the 2.6th percentile, MVRV-Z at the 23rd, long-term-holder NUPL at the 18.7th and only 66.4% of UTXOs in profit (28th percentile) mean the marginal seller below 60,000 is selling into the deepest-value band of the last several years. That argues the drawdown is shallow and slow rather than absent.

low

DVOL 34.66 sits at the 2.7th percentile of its 1-year range and 30-day ATM implied vol of 31.2% is at the 1.1st percentile of its 90-day range, while realized vol compresses across every window: 28.45% at 7 days, 30.88% at 30, 35.5% at 90. Every available measure points to the low end. That is the hazard rather than the comfort — implied volatility is priced for a quiet week directly into July payrolls and a Senate cloture window falling on the same day.