The same questions, answered against a week that came from the long side: bitcoin closed -3.2% at $83,679, $546.89m of positions were flushed in a day, and the funding clock sat flat at 0.002% throughout - the clearest evidence that this was a positioning event rather than a repricing of leverage.
This column answers the same questions every week - market structure, funding and basis, venues, liquidations and risk, outlook - so the series stays comparable week to week. This week every answer bends around one fact: the market absorbed a hawkish minutes release, a 30-year yield at a level last seen in 2002, and a $546.89m liquidation window without the funding clock moving at all.
Two dated prints framed the week. The September FOMC minutes on 7 October showed that most officials judged a further increase appropriate before year-end, and the 2026 median dot rose to 4.1%; that left October hike odds near 17% against 75-85% odds of a move by year-end. Then the long end repriced: the 30-year briefly touched 5.70% on 7 October, its highest since 2002, the 10-year closed at 5.24% (the 10-year held near 5.24% into the weekend) and the 2-year at 4.80%. Brent rose back above the round number on Gulf shipping risk and settled at $104.72 (+2.4%); WTI closed at $91.85 (+0.8%); gold added +1.2% to $4,220.30. Trump said on 8 October he would not attack Iran before the 3 November midterms.
| Reading | Note | Source | |
|---|---|---|---|
| US 10-year Treasury yield | 5.24% | held near 5.24% as the long end stayed elevated | Schwab Market Edge daily recap; StreetStats weekly, 9 October 2026 |
| US 2-year Treasury yield | 4.80% | -2bp on the week | StreetStats weekly, 9 October 2026 |
| S&P 500 | 7,811.54 | +1.2% on the week from 7,722.72 | Schwab Market Edge; IGM Financial, 9 October 2026 |
| Nasdaq Composite | 27,366.17 | +0.6% on the week | Schwab Market Edge; IGM Financial, 9 October 2026 |
| Dow Jones Industrial Average | 51,654.95 | +0.9% on the week | Schwab Market Edge; IGM Financial, 9 October 2026 |
| Gold | $4,220.30 | +1.2% on the week from $4,172.10 | COMEX settlement via Xinhua Finance and The National, 9 October 2026 |
| WTI crude | $91.85 | November contract, +0.8% on the week | NYMEX settlement via Xinhua Finance and The National, 9 October 2026 |
| Brent crude | $104.72 | December contract, +2.4% on the week | ICE settlement via Xinhua Finance and The National, 9 October 2026 |
| Crypto Fear & Greed | 56 | Neutral on the CoinMarketCap gauge; alternative.me read 61 | TokenPost and alternative.me, 11 October 2026 |
| Total crypto market cap | $2.80 trillion | bitcoin at $1.67 trillion of it, 59.5% dominance | CoinMarketCap aggregates via CoinPulseHQ, 11 October 2026 |
| US consumer sentiment, October preliminary | 46.3 | down from 48.1; released 9 October | University of Michigan via Schwab Market Edge, 9 October 2026 |
| US ISM services, September | 54.9 | prices-paid index 74.0; released 5 October | ISM via PANews, 5 October 2026 |
The cross-asset read is the week's most underrated signal. Equities rose - the S&P 500 +1.2%, the Nasdaq +0.6%, the Dow +0.9% - and gold rose +1.2% with crude. Only crypto fell. That combination, in a week when the VIX settled at 14.84, says the selling was asset-specific rather than a broad risk-off move: the market was not fleeing volatility, it was repricing crypto's own demand.
The week's structural read did not come from price. Hyperliquid's open positions reached a record 435,564 on 6 October, 115% above the 202,943 recorded in January, across 320,600 active wallets against 155,000, or 1.36 positions each. Open interest held at $17.29bn, $920m below the prior record, against a venue-mark count of $12.77bn. Daily volume, at $4.54bn, ran a 65% below the $12bn-$16bn August peaks.
| Metric | Reading | Note |
|---|---|---|
| Open positions | 435,564 | a record; 115% from 202,943 in January |
| Active wallets | 320,600 | from 155,000 in January; 1.36 positions each |
| Open interest, tracker cohort | $17.29bn | $920m off the prior record |
| Open interest, venue mark | $12.77bn | the venue's own two-sided count, printed alongside the tracker cohort |
| Daily perp volume | $4.54bn | inside a $2-6bn band; 65% below August peaks |
| 30-day perp DEX share | 32.9% | $202.6bn of $616.2bn cohort volume |
| 30-day fees | $87.5m | cumulative perp volume $5.4tn |
The important detail is the ratio, not the level. Open interest at close to four times daily volume means the book's exit door is narrow relative to the size standing in front of it. That is not automatically fragile - positions per wallet are unchanged at 1.36, so the growth is broad rather than concentrated in a few large accounts - but it does mean that when one cohort decides to leave at once, the flush is larger than the price move would imply. That is precisely what happened on 7 October.
The tell was what did not happen. A genuine leverage event flips funding sharply negative and steepens the curve; neither occurred. Binance's BTCUSDT funding sat near 0.002% per eight hours through the whole flush, and the dated curve held near 5% annualised, with the major December contracts at roughly 4.7-5.7%. The mid-week funding excursion was mild enough that longs briefly paid nothing rather than paying up.
What did move was the spot price of bitcoin in the United States relative to offshore markets. The Coinbase premium stayed negative for five straight sessions and widened to -0.118% by Sunday - a persistent U.S. discount. That is the single cleanest read of the week: the absence of a leverage premium plus the presence of a spot discount equals a sell-off driven by spot supply, not by borrowed supply.
| Signal | Reading | Interpretation | Source |
|---|---|---|---|
| Perpetual funding, 8-hour clock | 0.002% | Binance's BTCUSDT funding sat near 0.002% per eight hours | Binance via FinanceFeeds |
| Coinbase premium | -0.118% | the U.S. spot discount stayed open all five sessions, widening to -0.118% on 11 October | TokenPost |
| Dated-futures basis | about 5% annualised | the dated curve held near 5% annualised, with the major December contracts at roughly 4.7-5.7% | Zerocap weekly wrap |
| Wednesday flush, peak hour | $412.62m | $400.24m of longs (97.0%); shorts only $12.38m | CoinGlass via TokenPost |
| ETF weekly flow | -$681.1m | the -$487.07m session mid-week | SoSoValue via Hodl.Press / TokenPost |
| Hyperliquid OI, tracker cohort | $17.29bn | $920m off the prior record | HyperTracker via Cryptopolitan |
The US spot bitcoin ETF ledger is the best available read on whether new dollars are entering the collateral pool without borrowing, and this week it turned. Five sessions ran -$89.9m, +$118.86m, -$487.07m, -$244.13m, +$21.13m - a net $681.1m - and the three-week streak that had built $3.08bn ended. three straight positive weeks had lifted the year-to-date ledger before this week's reversal. The year-to-date ledger still reads about $494m, but its buffer is now thin.
The composition matters as much as the total. Ether ETFs shed $542.07m (a ninth straight outflow session) while NEAR funds took $35.99m, XRP funds $11.31m and SOL funds gave back $24.81m. The rotation was not into another large-cap crypto wrapper; it was small-cap specific and, in aggregate, out. That is consistent with the dominance move - bitcoin's share rose to 59.5% (up from 59.3%) - because when money leaves the complex, the largest asset falls least.
| Session | Net flow | Note |
|---|---|---|
| 5 October | -89.90m | the week opened negative |
| 6 October | +118.86m | a brief rebound ahead of the Fed minutes |
| 7 October | -487.07m | the week's worst session of the five |
| 8 October | -244.13m | selling continued as oil and yields rose |
| 9 October | +21.13m | a tentative inflow to end the week |
| Five-day total | -681.11m | a net outflow that ended a three-week inflow streak |
The on-chain cohort's centre of gravity is unchanged. Hyperliquid took 32.9% of the $616.2bn trailing-30-day perp DEX volume, printed $87.5m of 30-day fees, and has settled $5.4tn of cumulative perp volume. Its own open interest was $17.29bn on the tracker cohort and $12.77bn on the venue's two-sided mark - two measures of the same book, printed side by side because they answer different questions. The composition data - 435,564 positions across 320,600 wallets - comes from the venue's own tracker cohort, and the desk prints it with its name and snapshot time every week.
Printing both is the desk's method for not letting a single point travel further than its definition allows: the tracker cohort counts every open position the venue's analytics layer can see, the two-sided mark is the venue's own aggregate of the same book, and a third-party cohort read puts the wider perp-DEX complex at $14.80bn of open interest against $616.20bn of trailing 30-day volume, of which Hyperliquid is $202.6bn across 178 markets. The on-chain book is now large enough to set the week's tone and still small enough that a single cohort's exit clears an hour of turnover - the condition the 7 October flush made concrete.
The flush was the week's risk event. $546.89m liquidated over 24 hours on 7 October after a third failure at the round number; the peak hour took $412.62m, $400.24m of it longs (97.0%), with shorts just $12.38m and ether positions contributing $174.93m across the window. about 101,772 traders were closed out. The risk lesson is symmetrical to September's: two weeks ago the same machinery closed $750.5m of shorts; this week it closed $479.73m of longs. Stop placement above the round number was the trade's weak point both times.
The microstructure is the part that mattered. The trigger was a slide through $83,800, a level the market had defended twice; once it broke, the positions stacked above it unwound in a single hour that was 97.0% long-side, and $479.73m of the day's $546.89m was the long cohort's. open interest across the 21 exchanges CoinGlass tracks rose from $54.2bn to $55.3bn in the six hours after the flush. That is the signature of forced derisking rather than a directional repricing: open interest was rebuilt within hours, so the capital that left was replaced rather than lost. The split was $142.87m bitcoin and $174.93m ether, which tracks the majors' relative depth more than any single narrative.
Funding rates mean nothing without a measure of the collateral that can be posted against them, and this week that measure shrank. Total stablecoin supply fell to $304.2bn, supply contracted about $0.9bn on the week, with USDC leading the seven-day decline. A falling collateral base alongside a flat funding clock is a specific configuration: the leverage that exists is being maintained rather than expanded, and there is less dry powder behind it. When the slow variable falls while the fast variable holds, the book is trimming its capacity to add, not its willingness to hold.
The composition reinforces the reading. The largest issuer remains the settlement layer and the dominant form of posted collateral, so a contraction led by the second-largest issuer removes the marginal deposit rather than the base of the market. A supply that falls while funding holds and positions sit at a record describes a book funding itself from stock rather than flow: leverage is being maintained by rolling existing collateral, not by new deposits. That distinction is invisible in the funding rate and visible only in the stablecoin ledger, which is why the desk prints it every week.
The dispersion was the message. Bitcoin fell -3.2%, ether -8.3%, XRP -8.2%, ZEC -7.9%, and Solana -11.9% - the higher the beta, the deeper the drawdown, which is the opposite of a risk-on rotation. ZEC, September's leader, kept giving back its run at $1,237 (-7.9%) from $1,343. HYPE fell -4.2% to $86.72, 11.5% below its $97.98 all-time high, even as Hyperliquid's own book set a position record. The breadth was defensive: dominance rose to 59.5% (up from 59.3%). That is a macro-handoff tape in reverse - capital retreating to the largest asset rather than advancing into the smaller ones - and it is why the desk weights the spot ledger and the premium above alt perp activity in the read.
Four items landed inside the same seven days and they point the same direction. The CFTC proposed Regulation CTX and Regulation CAM, an advance notice of proposed rulemaking for leveraged, margined or financed retail crypto transactions; Michael Selig framed the package as a break from enforcement-led rulemaking. FinCEN withdrew both of its self-custody proposals - FinCEN withdrew both its unhosted-wallet recordkeeping rule and its crypto-mixing designation proposal on 5 and 6 October. The Treasury and IRS updated the staking safe harbour for crypto product trusts, and ESMA told EU firms to stop servicing stablecoins that do not comply with MiCA. More structure around leverage, less surveillance around self-custody: for a desk that prices perpetuals, the first half is the half that matters, because a federal registration path is what turns an offshore perp venue into a counterparty with a rulebook.
The broader tape makes the crypto-specific weakness look selective. The 30-year yield briefly touched a level last seen in 2002 (the 30-year briefly touched 5.70% on 7 October, its highest since 2002) and the 10-year closed at 5.24%; gold rose +1.2% to $4,220.30 and crude rose on Gulf shipping risk. Equities rose across the board: the S&P 500 +1.2% to 7,811.54, the Nasdaq +0.6% to 27,366.17, the Dow +0.9%, with the VIX settling at 14.84.
Bitcoin closed -3.2% in that environment. An asset that falls while equities, gold and oil rise is not trading a macro impulse - it is trading its own flow. The evidence is specific and repeated: an ETF ledger that turned negative at $681.1m, a Coinbase premium negative for five straight sessions at -0.118%, and a collateral base that contracted $0.9bn. That is the cross-asset signature of a demand problem, not a risk-appetite problem, and it is the read the desk will keep until the premium turns.
roughly 17%, against 75-85% odds of a move by year-end. The September minutes published on 7 October showed that most officials judged a further increase appropriate before year-end (the 2026 median dot rose to 4.1%), so the repricing point is the 14 October CPI and the 27-28 October decision rather than the October meeting.
Neither, and that is the point. Hyperliquid's open positions hit a record 435,564 while daily volume stayed at $4.54bn - a 65% decline from August peaks - and funding never left 0.002%. The 7 October flush cleared $479.73m of longs, but open interest rebuilt within six hours: open interest across the 21 exchanges CoinGlass tracks rose from $54.2bn to $55.3bn in the six hours after the flush
$546.89m over 24 hours, $479.73m of it longs (88%). The peak hour took $412.62m - $400.24m longs (97.0%) against $12.38m of shorts - and about 101,772 traders were closed out.
Because it isolates the buyer. The premium stayed negative for five straight sessions and reached -0.118% by 11 October, meaning U.S. spot traded below offshore prices throughout. Funding, by contrast, sat flat at 0.002%. A spot discount alongside a flat funding clock says the selling came from spot holders, not from leveraged positions being forced out.
It means the leverage is broad rather than concentrated - 435,564 positions across 320,600 wallets, or 1.36 each, little changed in character since January - but it also means exit capacity is thin. Open interest at close to four times the daily volume is why a move of less than two percent produced a $546.89m flush.
Two. The CFTC's Regulation CTX and Regulation CAM proposal gives exchanges offering leveraged retail crypto trading a federal registration path, with 60 days of comment from Federal Register publication. And FinCEN's withdrawals - FinCEN withdrew both its unhosted-wallet recordkeeping rule and its crypto-mixing designation proposal on 5 and 6 October - narrow the surveillance perimeter around self-custody. The first adds structure where perpetuals live; the second removes it where they settle.
That the marginal buyer withdrew. The ledger shed $681.1m across -$89.9m, +$118.86m, -$487.07m, -$244.13m and +$21.13m, ending a three-week inflow streak of $3.08bn. The year-to-date total is about $494m. Spot demand is bitcoin-specific, and the alternatives diverged: ether ETFs shed $542.07m while NEAR funds took $35.99m.