Bitcoin 2026-08-01 daily brief — AI stance: bearish

Perpetual funding is soft, not stretched. Weighted by open interest across venues it annualizes to 5.86%, a little over half the ~11% a year that the standard 0.01% per eight hours implies — that baseline is neutral, so anything near 5.9% is below neutral, and calling it crowded would be wrong. The largest single venue reads 5.77% annualized, 0.09 percentage points from the cross-venue figure, which is negligible: there is no pocket of venue-specific crowding to point at. More telling is the two-week path — the weighted rate rose from about 2.2% annualized on 7/20 to a peak near 9.4% on 7/28 and back to 5.9%, never once touching the neutral baseline even at its high, while price fell 5.25%. Leveraged longs did add into the decline, but at a scale that leaves nothing to unwind. The practical consequence is that the downside case here is not a leverage story. If price breaks, it has to break on spot supply — which is precisely where the ETF and US-premium evidence points.

Market-wide futures open interest is $47.75B, down 6.5% from $51.06B on 7/21 while spot fell 5.25% — so coin-denominated open interest is nearly unchanged at roughly 761k BTC versus 771k. Positions are being marked down, not closed, which means the underwater long inventory from the 66,257 high is still on the books and represents supply that only appears if the range breaks. Options open interest tells the other half: it dropped from $35.99B on 7/31 to $25.26B on 8/1, a 30% / $10.7B decline as the July monthly expiry rolled off on Friday 7/31. That removes whatever pinning anchor was holding price near the strike cluster, and it lands with 30-day implied vol at the 8.9th percentile of the last 90 days and skew at the 76.7th. Cheap vol, no gamma anchor, no leverage cushion, downside protection bid — a configuration that does not cause a move but does nothing to absorb one.

On 8/1, long liquidations of $24.14M against short liquidations of $2.06M give a 0.09 ratio, and shorts were the lowest of the entire two-week window — no squeeze fuel on that side either. Across 14 sessions longs lost roughly $371M against roughly $250M for shorts, a 1.5:1 tilt consistent with a grinding downtrend rather than a break. Two concentrated flushes stand out, $71.68M on 7/27 and $71.29M on 7/31, both on down days. But total activity sits at the 16.16th percentile of the past year: this is a slow bleed of long positioning, not a capitulation event. No washout has occurred.

The Coinbase premium is -9.73bps, tagged neutral by the framework but sitting right at the practical bottom-decile threshold of -10bps. The trend matters far more than the snapshot and it is one-directional: the seven sessions through 8/1 average about -10.6bps against about -7.4bps for the prior seven, with three prints beyond the decile line — -12.33bps on 7/26, -10.63bps on 7/29 and -13.78bps on 7/30. The spread has been negative on every one of the last 13 sessions and has been widening. Offshore is leading, US-side buying is not showing up, and this is an entirely separate measurement from the ETF creation data that reaches the same conclusion. Two independent proxies for US institutional demand both read empty.

ALIGNED. Every macro channel that transmits within a week points the same way as the on-chain flow read: the 10-year at 4.68% and rising, a Fed with three dissents for a hike, and Brent back to $90.74 reigniting the inflation input. The single dissenting channel is risk appetite — VIX 15.99, S&P within 0.26% of its 14-day high, the dollar through 100 at 99.803 — and BTC conspicuously failed to monetize it, underperforming the S&P by 4.8 points from 7/21 to 7/31 while the dollar fell. That converts the apparent conflict into confirming evidence: when the friendliest available macro backdrop produces no bid, the constraint is BTC-specific supply, not macro.

The dominant story is a spot demand vacuum with a paper trail. US spot ETFs posted a $265.4M single-day outflow on 7/31 led by IBIT at $122.7M — matching the -4,220 BTC print in the flow data — and the 30-day average sits at -1,223 BTC per day, roughly 2.7x the ~450 BTC of daily miner issuance in the current halving epoch. On 7/31 alone excess absorption was -4,711 BTC. The week netted -$61.53M for BTC funds while Ethereum funds took in +$27.42M, so this is partly rotation, not just de-risking. Coinbase's Q2 print on 7/30 put a P&L receipt on it: revenue $1.22B against $1.29–1.31B expected, transaction revenue -21% quarter-over-quarter, total spot volume -25% quarter-over-quarter, and a $359.5M GAAP net loss. Miner economics are stressed but not capitulating — Puell at 0.8292 (30.9th percentile) and the hash ribbon ratio at 0.9888, below 1 but recovering from 0.9807 on 7/20, with HUT, RIOT and MARA all reporting 8/4–8/6. On regulation, the CLARITY Act is practically dead for the year: Thune said on 7/23 it will not reach the floor before the 8/10 recess, cloture needs 60 votes, and no motion is scheduled — a standing bid under the sector is about to expire quietly inside this horizon. Separately, the Coldcard incident drained ~1,000 BTC from about 1,196 wallets in 41 minutes; the on-chain fingerprint fits 7/31 specifically, where address outputs hit a two-week maximum of 731,006 and fees per transaction spiked to 938.75 sats, 1.7x the next-highest reading, on a merely middling transaction count of 610,890 — a fan-out signature. The record 8/1 transaction count of 734,264 at the 99.53rd percentile came with fees at only the 11th percentile, so it is uncongested throughput and is not demand-confirming in either direction. Cycle position is roughly 28 months past the April 2024 halving with about 20 months to the next.

Hawkish. Policy sits at 3.63% with the 10-year at 4.68%, and the July 28–29 FOMC held rates with three regional presidents — Hammack, Kashkari and Logan — dissenting in favour of a 25bp hike, the first three same-direction dissents since September 2016. The statement cited inflation above the 2% goal from Middle East supply shocks and tariffs, and Brent settled 7.9% higher at $90.74 on 7/29 after Trump threatened retaliation against Iran, which reloads exactly that channel. The one countervailing input is M2 running +5.53% year-over-year — money growth is not tight even where rhetoric is. September is described as finely balanced and hinged on the next two CPI prints. Live sentiment reads Extreme Fear at 25 on the Fear & Greed scale, though that print is dated 7/18 and is two weeks stale, so weight it as background rather than as a current reading.

The 10-year is 4.68%, up 13bp from 4.55% on 7/20 and near the year's highs, with the Treasury quarterly refunding announcement on 8/5 as the supply event inside the horizon. Rising long-end yields against a Fed with three hike dissents is a direct headwind for a zero-cash-flow, long-duration asset. There is no credit-spread feed in this dataset, so I have no read on credit conditions and am not going to infer one from equity vol.

The dollar index closed 7/31 at 99.803, down 1.67% from 101.502 on 7/27 and through the 100 line for the first time in the two-week window. That is normally a clean tailwind for BTC. It was not converted: over the same 7/27–7/31 stretch BTC fell 1.35% from 63,750 to 62,888. A falling dollar that BTC cannot monetize is more informative than the dollar move itself.

Risk-on. The S&P closed 7,489.72 on 7/31, within 0.26% of its 14-day high of 7,509.20 set on 7/21, with VIX at 15.99 after spiking to 20.66 on 7/29. BTC did the opposite over the identical span: from 7/21 to 7/31 the S&P lost 0.26% while BTC lost 5.08%, a roughly 4.8-point underperformance into a friendly equity tape. Crypto is not trading as a high-beta risk asset right now; it is trading its own flow problem.

With 30-day realized vol at 31.46% annualized, a one-standard-deviation week spans roughly ±4.4%, or about 60,050 to 65,520. That puts the 30-day low of 61,849 only 1.48% below spot — well inside one sigma, so a retest over the next week is close to a coin flip rather than a forecast. The 60-day low of 58,519 is 6.79% down, roughly 1.6 sigma, which is a single-digit probability under a normal distribution and meaningfully higher than that given BTC's fat tails and the fact that implied vol is compressed at the 6th percentile with realized vol having nowhere obvious to go but up. The real asymmetry sits below 61,849: coin-denominated open interest is flat at ~761k BTC while price has fallen 5.25%, so a break of the 30-day floor releases held long inventory into a market where liquidations have only reached the 16th percentile of the past year — no washout has cleared the way. Above, the ceiling is close and specific: 63,363 then 66,257, with the 200-day 13.6% overhead at 71,309, far enough that a genuine trend reversal is not a one-week proposition in either case.

low