Bitcoin 2026-08-10 daily brief — AI stance: neutral

Funding is cheap, not crowded, and that is the cleanest read in the book. Across venues weighted by open interest, longs are paying 6.84% annualized — below the roughly 11%/yr that the exchange-default 0.01% per 8h implies — so perpetual longs are being charged less than the structural baseline. The single major venue in the dataset shows 6.99% annualized, a difference of just 0.15 percentage points from the all-venue aggregate, so there is no venue-specific pocket of crowding to point at; the venues agree. What matters more is the fortnight: the open-interest-weighted rate ran from 9.4% annualized on 2 August down to 2.6% on 4 August and back to 6.84% now, never once above the neutral baseline. That means the entire 1-9 August advance from 62,780 to 65,029 — plus 3.58% — was bought with spot, not borrowed with leverage. There is neither a long squeeze to harvest nor a crowded position to unwind here.

Light, two-way, and shrinking. Market-wide futures open interest is $46.46bn, down 6.1% from the 5 August high of $49.50bn and 4.1% in a single session from $48.45bn on 9 August, against a price that barely moved over that stretch — position reduction, not fresh directional risk. Options open interest of $25.79bn looks collapsed against $35.99bn on 31 July, but that step down happened in one session (to $25.26bn on 1 August) and is the July expiry rolling off; since then it has been flat in a $25-27bn band, so read it as calendar mechanics rather than de-risking. Options volume of $1.07bn on 10 August against a 13-day average near $2.3bn says participation is thin. On the single venue tracked in isolation, open interest of $1.44bn — about 3% of the all-venue $46.46bn — has been remarkably stable in a $1.41-1.47bn band across every session of the fortnight, reinforcing that nothing is being built there either. The volatility surface adds the one asymmetry in the set: as of 11 August, 30-day ATM implied volatility of 33.12% sits in only the 11.1th percentile of its 90-day range while 25-delta skew of +0.0545 sits in the 70th percentile of its rolling 30-day range. The market is pricing calm and simultaneously paying up for downside protection.

The 10 August decline cleaned long-side leverage without triggering a cascade: $35.10m of longs against $6.38m of shorts, a ratio of 0.18 tagged MORE_LONGS_LIQUIDATED, with total liquidations at only the 29.3rd percentile of the past year. For scale, the -3.0% day on 31 July took out $71.29m of longs, so this was half that. The fortnight as a whole is two-way churn rather than one side building — shorts absorbed $37.95m on 5 August and $28.10m on 4 August, longs $24.14m on 1 August — which is what a genuine range looks like from the inside. Neither side has accumulated the kind of stranded position that funds a violent move.

US spot has not led once in the observable window. The Coinbase-versus-offshore spread is -8.59 bps today and was negative on all thirteen observed sessions, averaging -9.30 bps, with the widest at -13.78 bps on 30 July and further sub -10 bps prints on 29 July (-10.63) and 3 August (-11.09). Against a practical extreme threshold of ±10 bps and a p99 near ±15 bps, that is a fortnight spent at the edge of the offshore-leading zone — though the classifier still tags the current level NEUTRAL, and at -8.59 bps that tag is correct, so this is a persistence observation rather than an extreme reading. The only let-up was -4.87 bps on 8 August, which reverted immediately. The tell is the overlap: this negative spread ran straight through the strongest ETF inflow week since April ($853.5m, 3-7 August). US institutional demand is arriving through primary-market ETF creations without lifting the domestic spot offer, which makes that flow headline a narrower bid than it reads.

CONFLICT — and the side I take is BTC-specific supply over the supportive macro tape. Equities are at the top of their range with VIX at 15.46, gold is at $4,487, M2 is growing 5.53%, and September hike odds have fallen to 43.9%: on that backdrop BTC should be extending, and instead it managed +0.13% while the S&P did +5.97%. The local on-chain set (BOTTOM ZONE, 4/8 bottom triggers, reserve risk at the 2.34th percentile) says value; the local flow set says why value isn't converting — a treasury holder selling 1,690 BTC below cost, miners capitulating on hash, and thirteen straight sessions of no US spot premium. When a supportive macro tape produces no follow-through, the constraint is the supply side, and that argues against extrapolating the 1-9 August advance.

Three supply facts dominate the flow narrative. First, ETF demand genuinely improved and then broke: the 30-day average flow crossed positive to +212.32 BTC/day from -1,239.98 on 29 July, the 7-day average is +1,117.10 BTC, and the week to 7 August took $853.5m — the strongest since April, with IBIT at roughly $693m or 81% of it — but 10 August printed -1,422.77 BTC, the first negative day since 31 July, with excess absorption of -1,857.15 BTC against miner issuance (a flow-to-issuance ratio of -3.28) and a 30-day flow z-score of -0.57. Second, the treasury bid has become supply: Strategy sold 1,690 BTC between 3 and 9 August at an average $64,262, below its $75,419 cost basis, to fund STRC preferred repurchases and dividends, bringing 2026 sales to 6,948 BTC with no purchase since 22 June. Third, miners are stressed and distributing — hash ribbons at 0.9863 (8.93rd percentile, 30-day hash below 60-day), Puell 0.727, and fee income at the 10.96th percentile (476 sats/tx) despite transaction count in the 98.22nd percentile at 639,365. Offsetting: wallets holding 10 to 10,000 BTC added 0.34% of balance since 29 July, about 19,610 coins, while sub-0.01 BTC addresses trimmed 0.59% — large holders absorbing small-holder supply in the $62,000-$65,000 band. On catalysts, there is nothing inside the horizon: the CLARITY Act slipped to a 15 September cloture vote needing 60 votes, with the Senate back on 14 September. One tail item worth naming: an attacker has drained over $130m from Coldcard wallets since 30 July via a firmware flaw dating to March 2021, a marginal but real hit to self-custody confidence.

Hawkish, and the load-bearing point is what is actually being debated: the July 28-29 meeting held the target range at 3.50%-3.75% on a 9-3 vote with Hammack, Kashkari and Logan all dissenting in favour of a hike, and the live question into 15-16 September is hike-versus-hold, not cut-versus-hold — no cut is priced anywhere in the visible calendar. Effective fed funds sits at 3.63% with the 10-year at 4.65% and M2 growing 5.53% year-on-year, so liquidity is expanding while policy is still leaning against it. LSEG-implied odds did move dovish after the July payrolls contraction of 23,000 (versus roughly +80,000 expected, with May and June revised down a combined 103,000): the September hike probability fell to 43.9% from 57% and the hold reading rose to 60.4% from 43.2% as reported. That is a repricing within a hawkish setting, not a pivot. The live sentiment gauge is the weakest input here — Fear & Greed reads 25 ('Extreme Fear') but is stamped 18 July, roughly three weeks before this brief's data date, and VIX at 15.46 with the S&P at 7,753 flatly contradicts a fear reading, so it carries little weight. The 12 August CPI print (consensus 3.4% headline from 3.5%, core 2.5% from 2.6%) is the binding event inside the horizon; the 19 August FOMC minutes, the real read on that 9-3 split, lands just outside it.

The 10-year sits at 4.65%, four basis points off 4.69% on 9 August but above the 4.61% of 29 July, having traded a 4.61%-4.75% range over the fortnight — a high, rangebound long end roughly 102 bp above fed funds at 3.63%. This week carries heavy Treasury supply into that level, with 10-year note and 30-year bond auctions both at 1:00 p.m. ET, and auction reception is the flagged swing factor for yields and the dollar. There is no credit-spread feed in this dataset, so no investment-grade or high-yield spread read is available here; I am not going to infer one from the equity tape.

DXY 99.81, down 1.1% from 100.91 on 29 July but pinned in a 99.60-100.91 band all fortnight — soft without trending. For BTC at this stage the absence of a dollar breakdown matters more than the mild softness: the debasement bid that a sub-100 dollar plus 5.53% M2 growth should generate is visibly going elsewhere, with gold at $4,487, while BTC trades 21.23% below its 90-day high. BTC is not currently the vehicle for that trade.

Firmly risk-on and BTC is not participating — this is the single most important cross-asset fact in this brief. The S&P rose 5.97% from 7,316.15 on 29 July to 7,753.11 on 10 August, essentially the high of the window, with VIX falling from 20.66 to 15.46 over the same span. BTC went from 63,874.13 to 63,959.64 across those identical twelve sessions: +0.13%. A high-beta risk asset that captures 2% of a 6% equity rally is not being held back by macro conditions; the marginal buyer simply is not showing up.

Layered, and the near levels are close. The 50-day average at 63,338 is 0.98% away and is best treated as a coin flip inside the horizon given a 0.92% daily sigma — it is roughly one average day's move. Below that, the 30-day low of 61,849 is 3.41% down and sits about 1.4 sigma out on last week's realized volatility but only around 0.8 sigma on 30-day realized (29.77%) or 30-day implied (33.12%) — so if volatility reverts toward its own recent norm through CPI, PPI and retail sales, a test of that low is closer to even money than the calm surface suggests. Below the range, the 60-day low of 58,519 is 8.5% down; that requires a catalyst, and the plausible one is an upside CPI surprise repricing the September hike, not organic selling. Two things argue the downside is contained rather than cascading: total liquidations sit at the 29.3rd percentile of the year with funding at 6.84% annualized, so there is no leverage stack to unwind, and the biggest single day in the last thirty was only -3.0% on 31 July. Two things argue it is real: the largest holder in the corporate cohort is a seller below its own cost basis, and long-term-holder SOPR at 0.8896 shows coins already being spent at an average 11% loss, which is ongoing supply rather than a washed-out float. The 200-day average at 70,014 caps any recovery scenario and is 9.5% above spot — nothing in a 5-7 day window reaches it.

low

the daily dataset