Bitcoin 2026-08-24 daily brief — AI stance: neutral

Across venues, weighted by open interest, perpetual funding annualizes to about 3.1% — roughly a quarter of the ~11% that the standard 0.01% per-eight-hour default represents, so market-wide positioning is not merely un-crowded, it is running below the neutral baseline after a 25% advance. That is the single most important number in this brief and it cuts both ways: nobody is paying up to be long, so there is no leverage overhang to unwind, and equally there is no short base left to squeeze. It also fell sharply from yesterday, when the all-venue figure was running near 10.5% annualized, meaning leverage demand contracted while price made a new high — the hallmark of a spot-led rather than perp-led advance. One large venue is printing about 8.9% annualized, some 5.8 percentage points above the all-venue weighted figure; that tells us the residual long crowding is concentrated on that single venue rather than distributed across the market, and it is not a spread between two market-wide measures.

Market-wide futures open interest across all venues stands at $57.30bn, up 21.0% from $47.34bn on August 12 — a substantial build, but one that arrived alongside a 25% price advance, so open interest measured in coin terms has barely moved and leverage per dollar of price has fallen. That is confirmed by funding running below its neutral baseline. The more striking build is in options, where open interest rose 47.5% from $26.13bn to $38.55bn over the same twelve days and volume spiked to $11.98bn on August 22 before falling back to $2.71bn today. Combine that with 25-delta skew at the 10th percentile and 30-day ATM implied vol at 41.56%, the 84th percentile of its 90-day range, and the picture is a market that stopped expressing itself through leveraged directional futures and started buying convexity — specifically downside convexity — into a dated event cluster. That is not a topping signal; it is a hedged market, which is exactly the configuration that produces a range rather than either a melt-up or a cascade.

The pattern still favours the bulls, but with a fraction of the earlier force. Today saw $67.4m of shorts liquidated against $44.7m of longs, a 1.51 ratio, with total liquidations at the 78th percentile of the past year. Compare that with the squeeze itself: $811m of shorts on August 19, $262m on August 20, $470m on August 21. So shorts are still the ones being carried out at the highs, which argues against an imminent reversal, but the magnitude has collapsed by more than an order of magnitude from the peak — the fuel is gone. The other side of the ledger is quietly building: long liquidations of $114m on August 21, $62m on August 22, $43-45m since, are the largest sustained long-side stops of the fourteen-day window and they are happening while price grinds higher.

The US-versus-offshore spot spread is at +0.32bp, technically positive and the first positive print in two weeks. The level is neutral — well inside the ±10bp practical extreme — but the trend is the signal here, and it is a clean, monotonic fourteen-session compression: -10.73bp on August 12, -10.66bp on August 16, -8.50bp on August 18, -6.81bp on August 19, -5.01bp on August 20, -2.21bp on August 21, -1.59bp on August 22, -2.66bp on August 23, +0.32bp today. Domestic demand went from a deep discount to flat, which corroborates the $1.92bn ETF week from an independent measure and is the strongest single piece of evidence that the advance was not purely offshore leverage. It is also the reason the downside case here is a pullback rather than a retrace: the marginal US buyer is present, just no longer aggressive.

CONFLICT, and it is worth being explicit about which side wins where. The local data is constructive: all-venue funding is unlevered at ~3.1% annualized, MVRV-Z is only at the 39th percentile after a 25.3% advance, shorts were still being liquidated 1.51-to-1 at the highs, and the US spot premium flipped positive for the first time in two weeks. The macro tape is hawkish: yields rising to 4.74%, 28.5% odds on a September hike against 1% on a cut, sticky 3.3% core PCE landing tomorrow, and equities down while BTC ran. My resolution is not to split the difference — it is to assign each side the job the pipeline says it can do. The horse race puts macro only at 0.1704 average Sharpe, effectively no standalone directional content, while etf only at 0.873 is the only positive source group. So the flow and positioning data govern direction and they say the floor is solid; the macro calendar governs the volatility path and it says the ceiling is capped through August 28. That is a range, and it is a claim, not a hedge.

This advance has a clean, datable provenance and both of its engines are running down. Engine one was a squeeze: roughly $2.7bn of short positions were liquidated on August 19 as Trump convened a White House crypto meeting and pressed Congress on the CLARITY Act, producing the largest BTC advance since March. Engine two was real spot demand: US spot ETFs took in $1.92bn in the week to August 24, the strongest week since October 2025, with IBIT alone at $1.33bn over five consecutive days, and today's flow still absorbed 3.68x miner issuance with the 7-day average at the 85th percentile. Both are fading. The daily flow z-score has decayed from +2.47 on August 20 to +0.14 today, and August's $2.38bn — the strongest month of 2026 — still leaves the ETF complex about $2.91bn net negative year to date, so this is a recovery in flows, not a new secular bid. Against that, the regulatory pipeline is genuinely improved: the SEC proposed Regulation Crypto Assets on August 18 with a $5m startup exemption and a conditional safe harbour, and the Senate has a CLARITY Act procedural vote set for September 15 — though Galaxy Research cut its odds of passage this year from 50% to 30% over illicit-finance and stablecoin-rewards disputes, and September 15 is outside this horizon anyway. Miner economics are the weak leg: the hash-ribbons ratio at 0.986085 sits at the 9th percentile and has fallen every single day of the last fourteen, the Puell Multiple at 0.936159 is below one, and fee revenue is at the 11th percentile (493 sats per transaction) despite transaction count at the 98th percentile — near-record block demand generating almost no fee income. IREN reports Thursday with consensus revenue down 30% y/y and EPS of -$0.55 against +$0.65 a year ago, which is the same stress in listed form. The September 1 BIP-110 deadline is a non-event: the August 8 chain split stalled after two blocks with ~2.5% miner support against a 55% threshold.

Hawkish, and hardening. Fed funds sits at 3.63% while the 10-year is at 4.74% — 111bp above the policy rate and up 11bp from 4.63% on August 14, including +5bp today. Prediction-market pricing as of August 18 put September cut odds at roughly 1% against a 70.5% hold and a 28.5% quarter-point hike, so the live debate is hold-versus-tighten, not the timing of easing. Core PCE lands August 26 with consensus at +0.2% m/m and 3.3% y/y unchanged — sticky at a level that gives no cover for easing — and Kevin Warsh delivers his first Jackson Hole keynote as Chair on August 28. M2 is growing 5.53% y/y and gold is at $4,702.70, which says liquidity and debasement hedging are running well ahead of what a 3.63% policy rate would suggest, and is the strongest structural argument against reading this Fed as tight in real terms. On sentiment: the Fear & Greed reading of 25 (Extreme Fear) carries an as-of date of 2026-07-18, 37 days stale, and describes the pre-rally tape near 63,000 — it is not usable as a current gauge. The live sentiment reads are DVOL at 43.41 (55th percentile of the past year) and 25-delta 30-day skew at the 10th percentile of its 30-day range, i.e. puts bid into the event cluster.

The 10-year at 4.74% is rising into the event cluster: 4.63% on August 14, 4.69% yesterday, 4.74% today. That direction matters more than the level here, because the move is happening despite the Treasury's August 19 announcement that it will at least double long-dated buybacks from $2bn to $4bn per operation — an operation that runs September 9 to November 4, targeting the 10-to-20 and 20-to-30-year sectors, and therefore outside this brief's horizon. Yields firming ahead of a support programme that has not started yet is a hawkish signal, not a neutral one. There is no credit-spread feed in this dataset, so I am making no claim about credit conditions — that read is unavailable rather than benign.

DXY at 99.027, down 1.0% from 100.02 on August 12 but up from 98.821 yesterday — a soft dollar that has stopped softening. The August decline is the cleanest macro tailwind BTC had during this advance, and the two-day upturn into the PCE print and Jackson Hole is the first sign it is being handed back. At this stage — price at the 90-day high with valuation still at the 39th percentile on MVRV-Z — a flat-to-firmer dollar does not force a reversal, but it removes the follow-through that carried the move off 62,837.

Risk-on in equities is not confirming this move. The S&P 500 at 7,652.86 is down 1.87% from 7,798.99 on August 13 — the same window in which BTC rose 25.3% — and Friday's session was itself a bounce from a sell-off driven by a sharp rise in bond yields. VIX at 15.85 is low and unstressed, so this is not equity risk aversion; it is BTC decoupling upward on a crypto-specific catalyst while equities chop sideways-to-lower. Decouplings sourced from a single dated catalyst rather than a broad liquidity impulse historically resolve back toward the equity tape more often than they drag it along.

The asymmetry is unusually legible here because the levels are recent and the flow composition is known. The shallow case, and my base case, is a drift or drop back into the 76,998-72,661 shelf built between August 20 and August 22 — call it a 2% to 8% giveback — which would simply retrace the thinnest part of the squeeze without touching the valuation picture; MVRV-Z at the 39th percentile and Reserve Risk at the 11th mean nothing about the medium-term structure changes in that zone. The more serious case runs to the 200-day MA at 69,094, which sits almost exactly on the August 19 close of 69,221 and would be a 12.2% drawdown; that is the level I would expect to hold, because it is where the pre-catalyst price and the long-term trend reference coincide and where the ETF complex demonstrated real appetite. Below that, the 50-day MA at 65,451 is 16.9% down and the August 16 launch point at 62,837 is 20.2% down — reaching either would mean the entire regulatory-catalyst repricing was rejected, which requires a genuinely hawkish surprise rather than an in-line one. I would weight the shallow case as the clear favourite over a 5-7 day window, the 200-day test as a real but secondary possibility concentrated in the 48 hours after the August 28 keynote, and a full round-trip below 65,451 as a tail that needs both a hot PCE tomorrow and a hawkish Warsh. What raises all of these relative to a normal week is that implied vol is not compensating: 30-day ATM at 41.56% is below 7-day realized at 68.64%, so short-dated protection is cheap relative to what the market has actually been doing, and the 10th-percentile put skew says somebody has already noticed.

high