Bitcoin 2026-08-15 daily brief — AI stance: bearish

Perpetual funding is below neutral and has just reset hard — which directly contradicts the 'crowded longs' framing circulating in market commentary as recently as August 13. Weighting across venues by open interest, the annualized cost of holding a long is 2.87%, against roughly 11% for the exchange-default rate that represents genuine neutrality. Longs are paying almost nothing. That figure collapsed from around 9.8% annualized just one day earlier, so whatever long crowding existed mid-month was flushed out over August 13–15 and has not rebuilt. One major venue is running considerably hotter at 6.44% annualized, roughly 3.6 percentage points above the cross-venue aggregate, which locates the residual long positioning on that specific venue rather than in the market at large — but even that hotter venue reading is below the neutral baseline. Nobody is being punished for leverage in either direction. The practical implication for a bearish view is that any decline from here has to come from spot selling, not from a leverage cascade, which means a grind rather than a wick.

Market-wide futures open interest across all venues stands at $47.81 billion, down 3.4% from $49.50 billion on August 5 — leverage leaving the system alongside the funding reset, consistent with de-risking rather than a directional build. Options open interest is $26.26 billion against $2.88 billion of daily volume. The options surface is calm to the point of complacency: 30-day at-the-money implied volatility of 30.1% sits at the 1st percentile of its 90-day range, the 25-delta skew at 0.0355 is in the 23rd percentile of its 30-day range with no meaningful put bid, and the 30-to-90-day term structure is in contango at +0.0737, pricing more volatility later than now. Composite positioning: leverage cleared, no directional crowding on either side, and volatility priced at the floor. This is the market's least defended state — a directional resolution meets no positioning resistance in either direction, which raises the variance around my directional call without changing its sign.

August 15 saw $1.09 million of long and $0.91 million of short liquidations — a total at the 0.27th percentile of the past year, which is to say essentially nothing was liquidated. The 0.84 long-to-short ratio and the 'more longs liquidated' tag are noise on a near-zero base and should not be interpreted. The informative pattern is the three sessions before: August 13 saw $29.9 million of long against $8.2 million of short liquidations, and August 14 saw $21.1 million against $3.7 million. Longs were removed over those two days, funding reset immediately after, and then the market went silent. That sequence is a completed positioning reset, not an ongoing one.

The domestic-versus-offshore spot spread is -10.85 basis points, in the bottom decile of the ±10bp practical extreme band and inside a historical range of roughly ±24bp. Offshore venues are setting the marginal price. What matters more than the snapshot is the fortnight's shape, and it needs stating precisely: the spread has been negative on all fourteen sessions, recovered to a local best of -4.87bp on August 8, and has deteriorated steadily since to -10.85bp. It is not a new fortnight low — that was -11.09bp on August 3 — and it is only marginally below August 13's -10.76bp, so the last few sessions are better described as pinned near the lows than as breaking to new ones. The substantive read is that the domestic bid has been absent for two straight weeks and the mid-month attempt to recover it failed. That is the same channel showing up in the fund-flow data, seen from a different angle, and two independent measurements of a missing domestic bid is the most robust observation in this brief.

CONFLICT. The macro tape is unambiguously supportive — equities within 0.2% of record highs, equity volatility at 14.25, long rates 12bp lower on the month, inflation decelerating to 3.4%, a flat dollar. The Bitcoin tape is unambiguously not — below both moving averages, 18.68% off the 90-day high, the domestic bid gone, the long-term-holder cohort selling at a loss. I take the Bitcoin-specific side, and the divergence itself is why. An asset that will not rally into a favourable macro backdrop is telling you its problem is idiosyncratic — a demand deficit no amount of general risk appetite is currently reaching. The one caveat that cuts the other way: this same decoupling means a macro risk-off shock would find Bitcoin already discounted, so the downside is less levered to macro than usual.

Three supply-side deteriorations are compounding. First, the long-term-holder cohort is distributing: supply fell roughly 210,000 BTC from the July 29 peak of 16.82 million, the first weekly decline of 2026 and the largest two-week drop since December 2024, with that cohort's spending ratio printing 0.86, 0.90 and 0.86 across August 9–11 — coins moving 10–14% below cost — and 0.963 on the latest print. Second, the US fund complex flipped from buyer to seller: daily flow ran +3,784 BTC on August 5 and turned negative from August 10, with the seven-day average crossing from +1,834 BTC/day on August 7 to -336 BTC/day on August 14, and roughly $192 million exiting across the first back-to-back outflow days since late July. Excess absorption against miner issuance was -1,360 BTC on August 14, meaning the fund complex is no longer clearing new supply. Third, miner economics are compressing: the Puell multiple at 0.754 sits in the 24th percentile, the hash-ribbon ratio at 0.9887 in the 10th percentile has the 30-day hashrate average below the 60-day — the textbook miner-capitulation configuration — and fee revenue is 323.8 sats per transaction, the 10th percentile, while transaction count at 719,514 sits in the 99th. Record block-space usage generating bottom-decile fee income means miners are working harder for less, and the marginal miner sells. On regulation, MSCI's consultation could delete Bitcoin-treasury companies from its global indices — a simulation on May 2026 data removed Strategy, Metaplanet and Yellow Cake from ACWI IMI — with feedback open to September 30 and no effect before the November 2026 review. That is a forced-selling overhang on the corporate-treasury bid, dated well outside this horizon but structurally relevant. On cycle position, this is roughly 28 months past the April 2024 halving, historically the post-peak stretch, which sits awkwardly beside the bottom-zone valuation reading.

Hawkish. The funds rate is 3.63% against July CPI at 3.4% year-over-year, leaving roughly 20bp of positive real policy rate — technically restrictive, barely. What makes the stance hawkish is not the level but the posture: the Committee held at 3.50–3.75% on a 9-3 vote on July 28–29, and the direction of those three dissents has not been disclosed. Financial press is debating a September hike rather than a cut, with money supply growing 5.53% year-over-year and year-ahead consumer inflation expectations rising to 4.3%. July CPI at +0.1% monthly and cooler-than-expected PPI did not shift the debate. The August 19 minutes resolve the dissent direction and are the single highest-impact scheduled event inside this brief's horizon. On live sentiment, the Fear & Greed gauge reads 25 — Extreme Fear — but that print is dated July 18, roughly four weeks stale, and cannot be treated as describing the current tape.

The 10-year yield is 4.63%, down 12bp from 4.75% on August 3 and back to its early-August low. Against a 3.63% funds rate that leaves roughly 100bp of positive curve slope — steep, which prices term premium and inflation risk rather than imminent easing. A twelve-basis-point rally in long rates alongside cooling CPI is a mild tailwind for duration-sensitive risk assets, and Bitcoin has not taken it. There is no credit-spread feed in this dataset, so I have no read on high-yield or investment-grade spreads and will not manufacture one.

Flat and inert. The dollar index closed at 99.64 on August 14 and has traded a 99.60–100.02 band over the full fortnight — a 0.42% range around the 100 handle. There is no dollar impulse in either direction, so the currency channel is simply not a driver of Bitcoin here. Anyone attributing this month's price action to the dollar is fitting a narrative to a flat line.

Risk-on, emphatically, and that is the problem. The S&P closed 7,785.76 on August 14, 0.17% off its August 13 high of 7,798.99 and up 2.4% from 7,600.50 on August 3. Equity volatility collapsed from 16.50 on August 4 to 14.25 on August 14. Over that same nine-session window Bitcoin fell from roughly 63,550 to 62,830. Equities made highs and Bitcoin made lows simultaneously. That decoupling is the most informative macro fact available.

The near-term downside is shallow but the tail is not. First support is the 30-day low at 62,780, just 0.47% below spot, and it has held. Below that, the 62,000 handle and then a gap down to the 60-day and 90-day low at 58,519, 7.2% lower. At 7-day realized volatility of 14.15% annualized — roughly 0.75% per session — a five-to-seven-day move at current volatility is only about 2%, so the modal bearish outcome over this horizon is a drift to the low 62,000s rather than anything dramatic. That is the honest central case. The tail is a different question: 30-day implied volatility at the 1st percentile of its 90-day range and the Deribit index at the 5th percentile of the past year mean volatility is priced for continuation of calm, and volatility at a floor in an established downtrend historically resolves in the trend's direction with amplification. A break of 58,519 would open the range beneath it with no obvious structural support until materially lower. Weighing these: I would put roughly 60% on a lower close seven sessions out, most of that probability mass in a 1–3% drift, with perhaps a 10% tail on a volatility-expansion move exceeding 6%. The cleared leverage argues that a cascade is less likely than the volatility compression alone would suggest — there is nothing left to liquidate.

low