Market makers rebuild BTC basis trades as funding rebounds and on-chain shorts swell
CME bitcoin futures open interest rose to about 122,000 BTC as CryptoQuant flagged a rare hedge-fund net-long flip.
Bitcoin’s post-rally funding rebound is reviving market-neutral cash-and-carry trades, with large short perpetual positions visible on Hyperliquid. At the same time, CME positioning data points to an unusually directional turn among hedge funds, complicating the usual basis-trade read.
Key Takeaways
- Hyperliquid on-chain data tied to Abraxas Capital, Fasanara Capital, and Wintermute showed short positions totaling 138,569 ETH (~$338 million) and 3,425 BTC (~$265 million).
- Abraxas Capital withdrew 73,872 ETH (~$173 million) from Binance over the past four days, per Arkham Intelligence.
- BTC perpetual funding rebounded to 6.7% annualized on a 30-day average as of Aug. 24 and 8.7% annualized on a 7-day average, based on Aegis data.
- CME bitcoin futures open interest climbed from roughly 87,000 BTC to 122,000 BTC in recent weeks, while CryptoQuant flagged CME hedge funds as recently flipping net long.
Hyperliquid Shorts Point to a Basis-Trade Rebuild After BTC’s Sprint
Bitcoin’s sprint from roughly $62,000 to above $77,000 in a matter of days last week reset the derivatives tape. About $3 billion in losses hit leveraged short sellers as the move forced bearish positioning to cover. The second-order effect was cleaner: funding flipped firmly positive, and the carry desks came back.
On-chain positioning on Hyperliquid made that shift unusually visible. Lookonchain-tracked data linked Abraxas Capital, Fasanara Capital, and Wintermute to short perpetual futures positions totaling 138,569 ETH (about $338 million) and 3,425 BTC (about $265 million). Those are not small “views.” They are the kind of size that typically maps to a balance-sheet trade.
The market-neutral interpretation is straightforward. When funding turns positive, being short perps can become an income stream if the short is paired against spot. That is the classic cash-and-carry, or basis trade. The catch is the evidence here is positioning, not intent. The material does not confirm whether the Hyperliquid shorts are fully hedged with matching spot longs, or whether some slice is directional.
Arkham Intelligence data adds a clue but not a confirmation. It shows Abraxas Capital withdrew 73,872 ETH, worth about $173 million, from Binance over the past four days. That flow is consistent with building or relocating spot inventory, which is what a hedged carry book needs. It still does not prove the hedge ratio or where the spot is held.
CME Open Interest Jumps as CryptoQuant Flags a Rare Hedge-Fund Net-Long Flip
The same carry logic is showing up on regulated rails, where the basis trade has historically been easiest to warehouse at scale. Glassnode data shows CME bitcoin futures open interest rose from roughly 87,000 BTC to 122,000 BTC in recent weeks. That is a material increase in outstanding risk, and it lines up with the timing of the funding rebound.
What is unusual is the reported direction of one key cohort. CryptoQuant data shows hedge funds on CME recently flipped net long on bitcoin futures, described as rare because the basis trade structurally tends to keep that cohort net short. In the standard footprint, hedge funds are short futures against spot exposure elsewhere, harvesting the spread and keeping delta tight.
A net-long posture suggests at least some participants are not just clipping carry. They are leaning into upside through futures. That matters for market structure because it changes who is the marginal buyer on strength and who is forced to de-risk on weakness. A carry book can often sit through chop as long as funding stays paid. A directional long book has a different pain threshold.
The unresolved piece is duration and magnitude. The data point is described as a recent flip, but it is not quantified in size or persistence. Without that, it is hard to separate a tactical positioning shift from a structural change in how the cohort is running risk.
ETH Perps Re-Lever: $14B Open Interest and the Risk of Crowded Positioning
Leverage is not rebuilding only in bitcoin. Coinalyze data shows aggregated open interest in ETH perpetuals climbed to $14 billion, a level not seen in several months. That is the kind of number that changes liquidation dynamics even if spot is not moving much.
ETH perp leverage matters here because basis trades and directional trades often share the same plumbing. When open interest rises quickly, the market becomes more sensitive to funding swings and to short, sharp spot moves that force liquidations. That is true even for “market-neutral” books, because the hedge is never perfect in practice. Execution, collateral constraints, and venue-specific liquidation rules all create gaps.
Traders have started to flag that fragility. Trader @LLuciano_BTC said bitcoin has been “struggling to push meaningfully higher” even as funding turned positive, calling the setup “fragile.” That is the classic warning sign for a funding-led market: positioning gets bullish before price confirms.
Trader @misterrcrypto pointed to “crowded leveraged longs” building alongside the rally, raising the odds of a sharp flush or pullback if momentum stalls. That kind of move is not just a spot event. It is a funding event, an open-interest event, and a collateral event, all at once.
Funding Is Back in the High Single Digits: The Carry Math Traders Are Watching
Funding is the price of leverage in perpetual futures. Perps do not expire, so exchanges use periodic funding payments to keep perp prices anchored to spot. When funding is positive, longs pay shorts. When it is negative, shorts pay longs.
That sign flip is why the basis trade “turns on.” Aegis data put the 30-day average BTC perpetual funding rate at 6.7% annualized on Aug. 24, with the 7-day average at 8.7% annualized. Coinglass data cited BTC funding as positive and elevated across major exchanges at around 0.01% per 8 hours. Coinalyze cited aggregated perpetual funding rates around 0.0109% per hour for bitcoin and 0.0087% for ether.
For a carry desk, those are workable numbers. High single-digit annualized yield becomes meaningful when the book is hundreds of millions and the operational stack can keep basis tight and collateral efficient. 21Shares Capital Markets framed it bluntly: “the BTC basis remains rich, funding itself is sitting at what it characterized as standard levels — suggesting the trade is attractive but not yet overheated.”
The bigger issue is path dependency. Glassnode data indicated BTC annualized perpetual funding was compressed or negative for much of February through July as bitcoin sold off from all-time highs above $120,000 and leveraged longs were unwound. In that regime, carry dries up or turns into a cost. This month’s surge reversed it quickly, but it can reverse again just as fast.
My Read: The Market Is Paying Shorts Again—But the CME Long Signal Is the Wildcard
The threshold that matters is whether BTC funding holds near the recently cited 6.7% (30-day) and 8.7% (7-day) annualized levels from Aegis. If it does, the Hyperliquid short-perp size reads more like inventory management and yield capture than a wave of outright bearish conviction.
The real test is whether CME open interest can stay elevated above the roughly 122,000 BTC level from Glassnode while the CryptoQuant hedge-fund net-long posture persists. If that net-long signal sticks, the setup starts to look structural rather than narrative-driven because the marginal risk on CME is no longer just carry, it is direction.
