Bitcoin 2026-08-22 daily brief — AI stance: bearish

Cool, and this is the single most bullish structural fact available. Open-interest-weighted funding across exchanges annualizes to 12.5%, against the exchange-default 0.01% per eight hours that annualizes to roughly 11% — so the market-wide cost of carrying a long is barely above the neutral baseline after a 25% five-session advance. That is remarkable and it is not a crowded long; it says the move was driven by spot buying and short covering rather than leveraged chasing, which is why the bias reads neutral. One venue-specific note: that major venue prints 10.95% annualized, about 1.55 percentage points below the all-venue open-interest-weighted aggregate, so what modest crowding exists sits away from it rather than on it. Neither figure is anywhere near stretched, and describing this tape as having a leveraged long problem would be wrong.

Leverage rebuilt on the way up and is now trimming. Market-wide futures open interest stands at $55.32bn, having climbed 22.5% from $46.46bn on 08-10 to a $56.89bn peak on 08-21 before slipping 2.8%. That build is roughly in line with the price move, so leverage participated rather than led — consistent with funding staying at baseline. Options grew far faster than futures: options open interest is $39.26bn, up 52% from $25.79bn on 08-10, and options volume hit $11.98bn on 08-22 against $1.07bn on 08-10, roughly eleven times. That is a very large hedging and speculation impulse, and given where the skew sits, the hedging interpretation dominates. Net read: no leveraged long excess to purge, but an options complex that has bought a great deal of downside into strength.

A textbook squeeze that has finished. Short liquidations ran $811m on 08-19, $262m on 08-20 and $470m on 08-21, roughly $1.54bn in three sessions, as concentrated shorts built through months of 62,000 to 66,000 consolidation were forced out above 72,000. Reporting put the most recent 24-hour short-liquidation total near $1.06bn, more than 60% below the initial burst. On 08-22 the pattern reverted to longs-dominant, 62.4m against 26.2m for a 0.42 ratio — the first longs-dominant session since 08-16, which is a return to the pre-squeeze norm rather than a new signal. Total liquidation activity is at the 65th percentile of the past year, elevated but not exhausted. The implication is directional: the forced buyer is gone, so the next marginal bid has to be voluntary.

The trend is the signal and it is genuinely bullish. The US-versus-offshore spot spread stands at -1.59bps, well inside the neutral band, but it got there by improving in every single session for eight straight days — from -10.85bps on 08-15, which is essentially at the practical extreme threshold, through -10.66, -10.07, -8.50, -6.81, -5.01, -2.21 to -1.59. That is a complete repair of a deeply offshore-led, risk-off spot configuration, and it corroborates the ETF flow reversal from an independent measure. It is the cleanest counter-evidence to my directional call: US spot demand is no longer absent. What it is not yet is US leadership — the spread is still marginally negative, so offshore has not actually handed the bid over.

no field names quoted in prose above

CONFLICT. The local crypto data is bullish — flow reversal, funding at baseline, cycle percentiles in the 27th to 38th range, US spot spread repaired. The macro tape is not: equities down 1.4% to 2% on the week, the 10-year pinned near the top of its range at 4.69%, a hike rather than a cut as the live September question, oil firming into a PCE print, and an options market paying up for downside protection through the entire rally. BTC rose 25% while risk assets fell, on catalysts whose fundamental content is thinner than the move. I am siding with the macro tape for this horizon.

Three catalysts landed on 08-18 and 08-19 and the rally starts on 08-19. On 08-18 the SEC proposed Regulation Crypto Assets, its first bespoke crypto offering regime, with a $5m startup exemption and a $75m-per-12-months fundraising exemption plus a conditional safe harbor from investment-contract status. On 08-19 the President hosted roughly two dozen crypto and finance executives at the White House pushing the Clarity Act, H.R. 3633, which passed the House in 2025 and remains stalled in the Senate. The same session brought the Treasury buyback and the dollar break below 99. Flow confirmed it: US spot BTC ETFs took $606.3m on 08-20, the largest daily figure since May and a fourth consecutive inflow day, with IBIT at $503.0m or 83% of the category. Daily flows ran 4,631 / 2,931 / 7,472 / 8,344 / 3,920 BTC across 08-17 to 08-21 against net outflows the prior week, lifting the 7-day average to 3,477 BTC versus 1,164 on 30 days. Two caveats on the micro. The regulatory content is thin relative to a 25% move — offering exemptions are an issuer and altcoin story with essentially no near-term BTC cash-flow impact, and the Clarity Act is still stalled. And the on-chain economy does not corroborate: transaction count at the 99.1st percentile, 704,507, alongside fee per transaction at the 11th percentile, 437 sats — heavy traffic with no fee-paying urgency. Miners remain unhealed, with the hash-ribbons ratio at 0.987, the 9.3rd percentile, still below 1 and still declining.

Hawkish. Policy rate 3.63% with the 10-year at 4.69%, a 106bp positive spread, and the live September debate is a HIKE at roughly one-in-three odds — there is no cut in the distribution. Jackson Hole runs 08-27 to 08-29 with the Fed chair's keynote 08-28 at 10:00 ET, 19 days before the 09-16 decision, and the reporting consensus is that his record of short statements and explicit independence from market pricing argues against pre-committing. Offsetting on the liquidity side rather than the policy side: M2 is growing 5.53% year-over-year, and the Treasury said on 08-19 it will at least double liquidity-support buybacks of longer-dated coupons from $2bn to at least $4bn per operation, effective 09-09 through 11-04. On the live sentiment gauge I have to disqualify the input: the Fear and Greed reading of 25, Extreme Fear, is stamped 2026-07-18 — 36 days stale, from before a 25% advance. It cannot be read as current sentiment and I am not using it.

The 10-year is 4.69%, sitting at the top of a tight 4.63 to 4.72 two-week range and unchanged over the last two sessions. Direction is the story: 10- and 30-year yields reached their highest levels since 2025 and 2007 respectively early in the week on fiscal-sustainability worries and heavy government plus AI-related corporate issuance, then the 08-19 Treasury buyback announcement pulled the 10-year down 6bp to 4.647% and the 30-year down 9bp to 5.196%. There is no credit-spread feed in this data, so I have no read on credit and will not manufacture one — the rates channel here is a duration-stress-and-relief story, not a credit story.

Weakening, and the timing is the tell. The dollar index ran 100.02 on 08-12, broke below 99 on 08-19, and printed 98.84 on 08-21 — the most recent value available, since today's reading has not landed. That is a 1.3% decline, and the break below 99 lands on exactly the session BTC rose 6.92%. At this stage — price extended 18.71% above its 50-day average — a weaker dollar is a genuine tailwind but a second-order one; it supports the move without explaining a 25% week.

Risk-off in equities, which makes BTC's week an upward decoupling rather than a shared risk-on. The S&P closed 7,674 on 08-21, the most recent print available, down 1.4% on the week, with the Nasdaq down 2% on a semiconductor-led selloff hitting AI-linked names. VIX was only 15.13 on 08-21 in a 14.25 to 16.01 two-week range — so equities fell without any volatility event. BTC rose roughly 25% from the 08-16 close of 62,837 to the 08-21 high of 78,447 across the same window. Crypto-specific catalysts, not a risk-on tape, drove this.

Calibrate everything against a weekly one-sigma of roughly 9.8%, which is what 70.64% annualized seven-day realized implies. On that scale the first shelf at 72,661, the 08-20 close, is only 5.6% below spot — about 0.6 sigma, a routine week. The level that matters is the confluence at 69,007 and 69,221, where the 200-day average and the 08-19 close sit within 0.3% of each other: that is 10.4% down, about 1.06 sigma, so call it a roughly one-in-six left-tail week on a normal read and worse than that in practice, since BTC's downside tail is fatter than normal. Below that there is a long gap to the 50-day at 64,860, 15.8% down. The structural exposure worth naming is that air gap: a full retrace to the 200-day confluence never trips a 50-day trend gate, so a long-or-flat trend rule holds the entire 10.4% without a single exit signal. And the honest reason confidence here is low rather than medium: the 08-21 high at 78,447 is a mere 1.88% above spot, about 0.2 sigma, so the upside break is at least as reachable as the shallow support test. This view is being expressed inside a range narrower than one week of noise.

High, with an important divergence to name. Seven-day realized volatility is 70.64% annualized, running 1.69 times the 90-day at 41.71% and 1.58 times the 30-day at 44.85%, with three of the last four full sessions moving 4.85%, 6.92% and 7.66%. That is a high realized-vol tape by any reading. But implied has not followed: the 30-day implied-vol index sits at 42.27, only the 48.8th percentile of the past year, essentially its median, and 30-day at-the-money implied is 40.0%. So options are pricing roughly 40% forward against 70% delivered over the past week — the market is treating the volatility spike as transient and about to decay. On the implied gauge alone this would read moderate; I am calling it high because realized is what a position actually experiences, and the gap means downside protection is not expensively priced relative to what the tape is currently delivering.

strategies named only in key risk