The crossover happened quietly, in a week the headlines spent on the Fed. This note takes the arithmetic apart: what the 21-month crossover is made of, why ZEC is the cleanest laboratory for what comes next, and the case against the desk's own reading.
During the week to 6 September 2026, aggregate altcoin futures open interest overtook bitcoin's for the first time in 21 months. The claim comes from Zerocap's weekly wrap and is verifiable in shape against the pieces this site tracks independently: bitcoin's book at $53.44bn and -1.64% on the week, the aggregate flat near $136bn. The question is not whether the crossover happened but what it is made of - and what it predicts, if anything, for the liquidation mechanics that follow.
Three facts fix the anatomy. First, the total did not expand: $136bn in aggregate futures open interest, unchanged for a second consecutive day into 2 September. Second, the bitcoin side contracted: $890m left the bitcoin book on the week, and the intra-week high of $58.12bn - printed during the 4 September squeeze - was given back by the close. Third, the growth sat elsewhere, in a complex that had lost nearly 23% in the first half of 2026. This is relocation of risk, not creation of it.
| Fact | Reading | What it establishes |
|---|---|---|
| Aggregate futures OI | $136bn | unchanged into 2 September - relocation, not expansion |
| Bitcoin futures OI | $53.44bn | -1.64% ($890m) on the week |
| Altcoin aggregate | above bitcoin's for the first time in 21 months | growth returned first to a complex that lost nearly 23% in H1 2026 |
ZEC is the week's cleanest single-name case of what alt-side leverage does when it turns. The asset settled the week at $1,014.58, +22.21% higher, after touching $1,023.00 intraday on 4 September. On Hyperliquid its open interest rose to $542m from $485m a week earlier - +11.6% - and the venue's tracked cohort shows the asset funding at the standard 0.0013% per hour even as it moved 15.5% in a day. The forced-flow record: $36.6m of ZEC liquidated in 24 hours on 4 September, $34.5m of it shorts - 94% of the forced flow was bears being closed, not bulls. DASH (+30.52% on the week) and XMR (+16.03%) moved in the same pattern.
The instructive part is the funding rate. Through a 22% weekly move, the cost of holding the position stayed at the venue's average. The squeeze was financed by shorts closing, not by leveraged longs paying up for exposure - which means the positioning behind the move is less committed than the price chart implies.
| Measure | Reading |
|---|---|
| Weekly close | $1,014.58 (+22.21% on the week) |
| Intraday extreme, 4 September | $1,023.00 |
| Daily move at the snapshot | +16.55% |
| Hyperliquid open interest | $542m, from $485m a week earlier (+11.6%) |
| Aggregate futures open interest | $2.3-2.4bn |
| Liquidated in 24h, 4 September | $36.6m, of which $34.5m shorts |
| Funding at the snapshot | 0.0013% per hour - the venue's average |
Open interest is the fuel tank of the liquidation mechanism. A dollar of open interest is a dollar of positions that can be forced at a price, and the liquidation cascades that define crypto risk events run through whichever book is largest and most one-sided. A market where the altcoin book exceeds the bitcoin book - for the first time in 21 months, and after that book lost nearly a quarter of its value in the first half - has relocated its cascade surface to the names with thinner liquidity and wider spreads. The 4 September windows show the mechanism at work already: $557m forced in 24 hours, 84% of it shorts, at 99,000 traders.
| Window | Total | Long / short | Traders | Reported by |
|---|---|---|---|---|
| 24h to 3 September | $286.0m | $164m long / $122m short | 105,000 traders | Gate news, 3 September 2026 |
| 24h to 4 September | $557.0m | $88m long / $469m short | 99,000 traders | TASS |
| Around the payrolls print, 4 September | $278.0m | predominantly longs | - | Zerocap |
The crossover week forced both sides of the market inside 48 hours: longs on the 3 September risk-off day ($286m, 57% longs, largest single position $4.74m), shorts on the 4 September squeeze ($557m, 84% shorts), and longs again around the payrolls print (roughly $278m). The mechanism this note is about - cascades running through whichever book is largest and most one-sided - ran twice in three sessions, and the aggregate book still ended the week near $136bn. Relocation of the cascade surface, with the cascade already firing.
| Holding period | Funding cost | Share of notional |
|---|---|---|
| 1 day | $28.80 | 0.03% |
| 1 week | $201.60 | 0.20% |
| 1 month | $864.00 | 0.86% |
| 1 quarter | $2,628.00 | 2.63% |
The carry table is the quiet counterpoint to the liquidation record. Through a week that forced both sides of the market, the cost of simply holding exposure stayed at a few tens of dollars a day on $100,000 of notional - a fraction of one percent. Leverage was not priced out of the market by its cost; it was repositioned by fear. That distinction is why the desk expects the crossover to resolve on scheduled data rather than on funding pressure.
The crossover is not an alt-season verdict, and the desk wants three distinctions kept visible. It is a leverage-location signal, not a demand signal: nothing in it says spot buyers arrived. It is not confirmed by funding: the cohort-wide rate of 10.88% annualised on $9.87bn of open interest is the venue's own average, not a crowded-leverage print. And it is measured against a bitcoin book that was itself shrinking - the crossover can say as much about the denominator as the numerator.
The crossover week was also the two-narratives week. Governor Waller's condition - ""Give disinflation a chance. We can wait one meeting."" - cut September hike odds from 66.6% to 54.6% within hours and squeezed bitcoin through $82,000; the August payrolls print of 162k against 55k expected snapped them back to 60.4%. The 2-year yield fell 7bp to 4.30% on the remarks. Alt-side leverage built into this uncertainty rather than after it, which is either conviction that the Fed path is priced, or positioning that has not yet met its data surprise. The 11 September CPI decides which.
The desk's reading is that the crossover marks relocated, short-horizon risk that resolves through alt-side liquidation cascades. The strongest counter-case: the comparison base is distorted. Bitcoin's open interest includes large hedged and basis-trade positions whose coin-denominated book actually grew +3.2% on the week (709,000 BTC from 687,000); a shrinking dollar value on a growing coin book reflects price, not exit. If the numerator grew for similar hedging reasons, the crossover says less about directional altcoin risk than the raw reading implies. The funding evidence cuts both ways here: subdued funding is consistent with hedging demand, and also with squeeze mechanics.
The claim's weakest link is aggregation: "altcoin open interest" sums dozens of assets from sources with different coverage. The desk's verification path has three steps, each checkable from printed inputs. First, pin the denominator: bitcoin's aggregate book is directly observable at $53.44bn, and its coin-denominated composition (709,000 BTC from 687,000, +3.2%) separates exit from mark-down. Second, pin the total: the aggregate held near $136bn, so the altcoin side's size can be bounded from above without summing it. Third, spot-check the largest alt components - ether and the ZEC complex tracked in section 3 - whose open interest and funding print alongside price every snapshot. If the three legs hold, the crossover's direction is robust to aggregation noise even where its precise level is not.
Aggregate futures open interest on altcoins (assets excluding bitcoin) exceeded bitcoin's, per Zerocap's weekly wrap for the week to 6 September 2026. Bitcoin's own book closed the week at $53.44bn, down 1.64%, while the aggregate held near $136bn - relocation of risk, not expansion.
Because its price, positioning and forced-flow records are all independently verifiable for the same window: a 22.21% weekly gain to $1,014.58, Hyperliquid open interest up 11.6% on the week to $541.8m, and $36.6m liquidated in 24 hours on 4 September of which $34.5m - 94% - were shorts, all while funding stayed at the venue's average 0.0013% per hour.
The desk does not read it that way. Open interest measures leveraged positioning, not spot demand; funding stayed at the venue's own average through the move; and the crossover occurred against a shrinking bitcoin denominator. It is a leverage-location signal with liquidation-mechanics implications, not a demand verdict.
Persistence and funding. A second week with the crossover holding would move it from noise toward regime, and a sustained rise in OI-weighted funding above the cohort's own average would show committed rather than cyclical leverage. The 11 September CPI is the largest scheduled test.
Zerocap's weekly crypto market wrap for the week to 6 September 2026. The desk verified the direction of the claim against independently tracked bitcoin figures but not every component of the altcoin aggregate, and says so on the record in the note.
Roughly $2.3-2.4bn in aggregate futures open interest at the week's close, of which $542m sat on Hyperliquid - up +11.6% from $485m a week earlier. For a name that moved +16.55% at the snapshot and settled the week +22.21% higher, that positioning growth at average funding is the note's cleanest single-name exhibit.
The total aggregate futures book ended the week near $136bn, essentially unchanged. What changed is its composition: bitcoin's share contracted while the altcoin complex grew past it. No net leverage entered the market - existing risk simply changed address, from the deepest, most liquid book to thinner ones. The practical consequence is about where the next cascade runs, not how large it will be.