Skip to content
MCAP $2.92T ▼-2.62%
BTC $85,603 ▲+0.29%
ETH $2,696 ▼-0.15%
BNB $779.88 ▼-0.86%
XRP $1.500 ▲+0.49%
SOL $120.37 ▲+0.44%
DOGE $0.0950 ▲+0.47%
ADA $0.272 ▲+2.76%
TRX $0.3362 ▼-0.11%
LINK $13.91 ▲+0.16%
AVAX $11.50 ▲+5.38%
HYPE $91.54 ▼-1.92%
DOT $1.210 ▲+1.24%
Market structure · Advanced

The crypto basis trade explained: how funds earn the gap between spot ETFs and futures

How cash-and-carry between spot bitcoin ETFs and CME or offshore futures works, who runs it, how to read it in CFTC data, why it drives ETF flows, and what the October 2025 unwind taught.

Crypto Coin Show Editorial Desk·Updated October 6, 2026·21 min read·Educational, not investment advice

Key takeaways

  • The basis trade buys spot bitcoin (increasingly via spot ETFs) and sells an equal amount of futures to earn the gap; that gap approached 25% annualised on CME in February 2024 and had compressed to about 5% by April 2026, barely above a 4.5% Treasury bill.
  • CFTC Commitments of Traders data show the trade in plain sight: as of September 29, 2026, leveraged funds held 11,836 short CME bitcoin contracts against 4,980 longs, about 60% of a 19,596-contract open interest.
  • A large share of spot ETF “inflows” are hedged. Hedge funds held about 37% of spot bitcoin ETP assets in Q1 2025 13F filings, and their IBIT holdings fell 28% between Q3 and Q4 2025 as carry collapsed, which showed up as ETF outflows.
  • The offshore version, short perpetual swaps against spot, is where risk concentrates: on October 10, 2025, more than $19 billion of leveraged positions were liquidated in under 24 hours and Hyperliquid’s open interest fell 56% in one day.
  • In 2026 the trade rebuilt rather than disappeared: CME crypto futures went 24/7 on May 29, 2026, Hyperliquid open interest was back to $14.3 billion by September 8, 2026, and weighted perp funding read +8.4% annualised on October 6, 2026.

Who this is for: Fund managers, treasurers, analysts and serious investors who want to understand what the basis trade is, who runs it, how to read it in CFTC and exchange data, and how to size, cost and stress-test a cash-and-carry position before treating an ETF flow headline as a signal of real demand.

Every few months a headline announces record spot bitcoin ETF inflows, and a day later another asks why the price did not move. The answer is often the basis trade. Since US spot bitcoin ETFs launched in January 2024, a meaningful fraction of the money flowing into them has been one leg of a hedged position: long the ETF, short a CME futures contract of the same size, with the manager indifferent to where bitcoin goes and interested only in the spread between the two prices.

That spread, the basis, is a quiet engine of crypto market structure. When it is wide, hedge funds pile in, CME open interest rises and ETF creations surge. When it compresses, the same funds unwind, and the unwind looks exactly like a bear market in the flow data even though nobody changed their view of bitcoin. The Block reported on April 9, 2026 that CME bitcoin futures open interest had fallen to $7.2 billion, its lowest since February 2024, with the basis near 5% against a roughly 4.5% risk-free rate; a Two Prime executive quoted by DL News on April 20, 2026 called it a basis unwind masquerading as a bear market.

This guide covers the mechanics, the CME and offshore versions, who runs the trade, how to read the positioning data, how the basis moved from 2024 through 2026, and a worked example of a fund deploying $50 million.

The crypto basis trade by the numbers

~5%Annualised CME bitcoin basis, vs ~4.5% T-billsThe Block, April 2026
11,836Leveraged fund short CME BTC contracts (vs 4,980 long)CFTC COT via Tradingster, Sep 29, 2026
$7.2bnCME BTC futures open interest, 14-month lowThe Block, April 2026
37%Share of spot BTC ETP assets held by hedge fundsBlockworks, Q1 2025 13F filings
$19bnLeveraged positions liquidated in under 24 hoursFTI Consulting, Oct 10, 2025
+8.4%Weighted BTC perp funding, annualisedDefiRate, Oct 6, 2026

What the basis is and why it exists

The basis is the difference between a futures price and the spot price of the asset it references. Futures above spot is contango, a positive basis; futures below spot is backwardation, a negative basis. For a dated contract the basis shrinks to zero at expiry because the contract settles to a spot reference. CME bitcoin futures cash-settle to the CME CF Bitcoin Reference Rate, the same benchmark family many spot ETFs use for net asset value, which is why the ETF-versus-CME pairing tracks so cleanly.

In theory the basis should equal the cost of carry: the interest forgone by holding the asset, less any yield it pays. Bitcoin pays none, so its annualised basis should sit near short-term rates, about 4% to 5% in 2025 and 2026. In practice crypto futures have traded far richer for long stretches. CF Benchmarks recorded the CME basis approaching 25% annualised in February 2024 and exceeding 20% in November 2024, and CME’s OpenMarkets piece of November 5, 2025 noted SOL and XRP bases spiking to 50% annualised in July 2025.

The premium persists because the natural buyers of crypto futures are leveraged speculators who want long exposure without holding coins, while the sellers are a smaller pool of arbitrageurs who must tie up capital, post margin and carry counterparty risk. The premium is the price of leverage; the basis trade collects it.

Basis versus funding

Offshore, most volume is in perpetual swaps, which never expire. Instead of convergence, a funding payment settled roughly every eight hours flows from longs to shorts when the perp trades above spot, and the reverse when below. An 8.4% annualised funding rate means a short collects about 8.4% a year in eight-hour slices while the rate stays positive. Our guide on perpetual futures, funding rates and liquidations covers the mechanics; the key difference is that funding can flip sign in hours, whereas a dated future locks the basis in on day one.

How cash-and-carry works, leg by leg

  1. Buy spot. The fund buys bitcoin, or shares in a spot ETF such as IBIT or FBTC. ETFs became the preferred spot leg for US funds after January 2024 because they sit in an ordinary prime brokerage account, can be financed like any equity, and sidestep the questions in our custody guide. The cost is a management fee of roughly 20 to 50 basis points a year.
  2. Sell futures of equal size. One standard CME bitcoin contract is 5 BTC; micro contracts of 0.1 BTC allow fine-tuning. The fund sells enough contracts to match the spot position, picking the maturity whose basis offers the best annualised return for the holding period.
  3. Post margin. The clearing member requires initial margin on the short, a large fraction of notional given bitcoin’s volatility. Treasury bills are acceptable collateral at CME, so margin keeps earning a cash yield, but it still ties up balance sheet.
  4. Earn the basis. As expiry approaches the futures price converges on spot. Directional gains and losses on the two legs cancel, leaving the original basis as profit.
  5. Roll or unwind. Before expiry the fund either closes both legs or rolls the short into the next contract. The roll re-prices the trade: a rich next-month premium locks in another quarter of carry; a collapsed or negative premium means selling the ETF, buying back the futures and going home.

What the fund is really earning

The headline basis is gross. Against it sit ETF fees, commissions, financing on any borrowed portion of the spot leg and the opportunity cost of margin. CF Benchmarks’ October 2025 analysis worked a $100 million example at a 10% basis, 5.3% SOFR and 7.3% borrowing cost and arrived at a 9.43% annualised return on equity. The hurdle is the T-bill yield plus the fund’s cost of capital and a risk premium, so a 5% basis against a 4.5% risk-free rate, as The Block described in April 2026, is a trade with an expected excess return under 1%.

Who runs the trade and on which venues

Hedge funds on CME

Multi-strategy and macro funds run the regulated version, long ETF and short CME, and appear in the CFTC’s Commitments of Traders report as “Leveraged Funds”; quarterly 13F filings name them. Blockworks reported on April 28, 2025 that hedge funds held about 37% of spot bitcoin ETP assets per Q1 2025 filings, and Hedgeweek put Brevan Howard’s IBIT stake at 37.9 million shares, about $2.6 billion, in June 2025. CF Benchmarks’ February 19, 2026 analysis tracked the retreat: aggregate hedge fund IBIT holdings fell from 114 million shares in Q3 2025 to 82 million in Q4 2025, with Brevan Howard down 85% to 5.5 million shares and Farallon down 70%. These were not bitcoin bears; they were carry traders whose carry went away.

Market makers and authorised participants

Firms such as Jane Street, which Hedgeweek reported holding $1.46 billion of IBIT in mid-2025, sit inside ETF creation and redemption. An authorised participant hedges the bitcoin it must deliver with futures or perps until settlement, and may hold the hedged inventory for carry when the basis is attractive. The ETF plumbing and the futures market are one machine seen from two sides.

Crypto-native funds and protocols on perps

Offshore, the trade runs against perpetuals on Binance, OKX, Bybit, Deribit and Hyperliquid. The most institutionalised version is Ethena’s USDe, a synthetic dollar backed by spot crypto hedged with short perps, whose yield is the funding rate. CCS reported USDe supply peaking at $14.8 billion on October 4, 2025 and near $4.9 billion of crypto-backed supply on September 25, 2026, when Ethena added a tokenised-equity basis trade on Binance that had averaged about 11% annualised over six months before compressing toward 7%. Not every large spot holder hedges: bitcoin treasury companies such as Strategy are outright long, and DL News noted Strategy buying 24,761 BTC between April 6 and 13, 2026 while hedged funds were selling into a falling basis.

Venue and instrument Who clears / holds collateral Margin and collateral Basis or funding, latest reading Open interest, latest reading
CME bitcoin futures (5 BTC), cash-settled to BRR, 24/7 since May 29, 2026 CME Clearing via FCM, CFTC-regulated Initial margin a large fraction of notional; T-bills accepted About 5% annualised (The Block, Apr 2026) 19,596 contracts, about 98,000 BTC (CFTC COT, Sep 29, 2026)
Binance BTC perpetual Exchange itself, offshore Crypto and stablecoin collateral, high leverage Positive; not in the Oct 6, 2026 DefiRate sample About 125,000 BTC, $11.2bn (CoinDesk via CoinGlass, Dec 22, 2025)
Hyperliquid BTC perpetual (on-chain) Protocol contracts; HLP vault backstop; ADL USDC collateral +11.0% annualised (DefiRate, Oct 6, 2026) $3.3bn BTC perp; $14.3bn platform total (DefiRate Oct 6, 2026; The Block Sep 8, 2026)
OKX BTC perpetual Exchange itself, offshore Crypto and stablecoin collateral +7.2% annualised (DefiRate, Oct 6, 2026) $2.5bn (DefiRate, Oct 6, 2026)
Deribit BTC perpetual Exchange (Coinbase-owned) BTC and USDC collateral +2.1% annualised (DefiRate, Oct 6, 2026) $821m (DefiRate, Oct 6, 2026)

Reading the trade in CFTC and exchange data

The CFTC’s Commitments of Traders report, published each Friday for positions as of the prior Tuesday, splits CME open interest into Dealers, Asset Managers, Leveraged Funds, Other Reportables and Non-reportables. The September 29, 2026 report, as compiled by Tradingster, showed bitcoin futures open interest of 19,596 contracts. Leveraged Funds were long 4,980 and short 11,836; Asset Managers were long 5,069 and short 1,483; Dealers were long 6,892 and short 4,212. The leveraged fund net short of roughly 6,850 contracts, about 34,000 BTC, is the visible futures leg of the ETF basis trade. TFTC’s read of the September 8, 2026 data made the same point: gross longs and gross shorts expanded together, the signature of hedging rather than a bearish bet.

  • Leveraged fund shorts are not sentiment. The “record hedge fund short” headlines of November 2024, when CME open interest hit a record near 45,000 contracts per CF Benchmarks, coincided with a basis above 20%. The record short was a record carry trade.
  • Open interest follows the basis with a lag. CF Benchmarks tracked CME open interest from about 30,000 contracts in January 2024 to the November 2024 record and back to the low 30,000s by May 2025 after the basis fell below 10%; The Block recorded five straight monthly declines into the April 2026 low.
  • Compare CME with offshore. CoinDesk reported on December 22, 2025 that Binance’s bitcoin futures open interest (about 125,000 BTC) had overtaken CME’s (about 123,000 BTC, down from 175,000 at the start of 2025) for the first time since November 2023. Regulated OI falling while offshore OI rises means carry capital is migrating to the cheaper, riskier venue or leaving.
  • Watch the 13Fs. They lag 45 days but name holders. Q2 2026 filings reported by The Block on August 15, 2026 showed Harvard at 3,044,612 IBIT shares after a 43% cut in Q1 2026, Mubadala at 14.7 million and JPMorgan up to 10.4 million. Rising bank and adviser holdings alongside falling hedge fund holdings describe an ETF base shifting from hedged to unhedged.

CCS tracks flows daily on the ETF tracker and follows COT data across assets, for example the September 23, 2026 report on CME leveraged funds cutting 46.3 million XRP of shorts in a week.

How the basis moved from 2024 to 2026

2024, the boom. The ETF launch gave US funds a clean spot leg and the basis approached 25% annualised in February 2024 per CF Benchmarks. CME open interest rose from about 30,000 contracts to a record near 45,000 by late November 2024, when the basis again exceeded 20%. CF Benchmarks’ February 2026 retrospective described carry that regularly touched 15% to 20%.

2025, compression and reload. Carry trades eat their own returns. CF Benchmarks recorded the basis briefly below zero in March 2025 and under 10% in April 2025, and sFOX documented about $2.7 billion of IBIT outflows over six weeks attributed to arbitrage capital exiting. Then the trade reloaded: Blockworks recorded $3.4 billion of weekly inflows in the week of April 21 to 25, 2025, with Bloomberg Intelligence suggesting basis activity had resumed while Bitwise argued the money was long-term. By mid-year the trade was multi-asset: CF Benchmarks tracked CME ether leveraged fund net shorts growing from about 6,000 to 19,200 contracts between April and August 2025 against roughly $11.2 billion of spot ether ETP inflows.

October 2025, the deleveraging. FTI Consulting documented perp funding climbing from about 10% annualised to nearly 30% by October 6, 2025. On October 10, after a tariff shock, more than $19 billion of leveraged positions were liquidated in roughly a day across about 1.6 million accounts. The Defiant recorded Hyperliquid open interest falling from $14.7 billion to $6.5 billion, with auto-deleveraging forcibly closing profitable shorts. On Binance, USDe traded in the mid-$0.60s for about 40 minutes against $0.99 on Curve and Uniswap because the exchange marked collateral off its own thin order book; Binance later paid $283 million in compensation. 21Shares recorded nearly 90% of liquidations as longs and funding flipping sharply negative. Each failure hit basis traders directly: ADL removed hedges, collateral was mispriced, and funding briefly paid the wrong way.

Late 2025 to spring 2026, the unwind. CoinDesk put the CME basis near 5% in December 2025, down from a roughly 15% peak. Bitcoin fell from about $126,000 in October 2025 to about $67,000 by February 2026 per CF Benchmarks, and hedge funds cut IBIT holdings 28% in Q4 2025. The Block reported CME monthly volume of $163 billion in March 2026, down nearly 50% from the January 2025 peak, and open interest of $7.2 billion in early April. Crypto Briefing noted about $4.5 billion of spot ETF redemptions in June 2026 and year-to-date flows still $1 billion negative in early September.

Mid-2026, the rebuild. CME launched 24/7 crypto futures and options trading on May 29, 2026, after $3 trillion of notional crypto volume in 2025. Spot ETFs drew $3.52 billion in August 2026 and $3.8 billion in the three weeks to September 5, lifting assets to $101.3 billion. Hyperliquid open interest was back to $14.3 billion on September 8, 2026. Bitcoin rallied more than 40% in Q3 2026 to about $87,000 per CoinDesk, and on October 6, 2026 DefiRate’s weighted perp funding read +8.4% annualised against a 30-day average of 5.2%. The carry is back, though not at 2024 levels.

Why many ETF inflows are hedged, and why it matters

When a hedge fund creates $100 million of IBIT and sells $100 million of CME futures, the ETF reports $100 million of inflows and the fund’s net bitcoin exposure is zero. The ETF’s bitcoin is real and leaves the market, but the futures supply equivalent synthetic exposure to someone else; the price effect is close to nothing, the effect on flow statistics is large.

Three consequences follow. Inflow headlines overstate demand when the basis is wide, and outflow headlines overstate capitulation when it compresses. Flows and CME positioning should be read together: inflows with a jump in leveraged shorts is a carry week, inflows with flat shorts is more likely directional. And the hedged share is falling: CF Benchmarks recorded IBIT advisory positions growing 145% from 38 million shares in Q4 2024 to 93 million in Q4 2025 while hedge funds cut. Money that will not leave when the basis compresses is a different kind of money.

How we got here: a timeline

CME lists bitcoin futures. Cash-settled 5 BTC contracts referencing the CME CF Bitcoin Reference Rate give US institutions a regulated futures leg.

US spot bitcoin ETFs launch. Funds gain a prime-brokerage-friendly spot leg; CME open interest is about 30,000 contracts.

Basis approaches 25% annualised. CF Benchmarks records the richest CME premium of the cycle.

Record CME open interest near 45,000 contracts. Basis exceeds 20% as the carry trade peaks.

Basis briefly dips below zero. Arbitrage capital exits; sFOX tracks about $2.7 billion of IBIT outflows over six weeks.

Carry reloads. Crypto ETPs take $3.4 billion in a week; Q1 13Fs show hedge funds holding about 37% of spot bitcoin ETP assets.

The October 10 deleveraging. More than $19 billion liquidated in a day, Hyperliquid OI down 56%, USDe at $0.65 on Binance, $283 million in Binance compensation.

Binance overtakes CME in bitcoin futures OI. CoinDesk records the CME basis near 5% and CME OI at about 123,000 BTC.

13Fs confirm the hedge fund exit. Aggregate hedge fund IBIT holdings fall 28% in Q4 2025 while adviser holdings rise 145% over the year.

CME OI hits a 14-month low of $7.2 billion. The Block records a basis of about 5% against a 4.5% risk-free rate.

CME goes 24/7. Crypto futures and options trade round the clock from May 29, 2026.

The trade rebuilds. Spot ETF assets reach $101.3 billion, Hyperliquid OI returns to $14.3 billion, and the September 29 COT shows leveraged funds short 11,836 contracts against 4,980 longs.

Funding back to 8.4% annualised. DefiRate’s weighted perp funding reads +8.4% on October 6, 2026 with bitcoin near $86,000.

Worked example: a $50 million BTC basis trade

A fund deploys $50 million of equity into the regulated version on October 6, 2026. Assumptions, to be checked against live quotes: spot bitcoin at $86,000 (CoinDesk, October 5, 2026); the CME contract expiring in 90 days trades at a 5% annualised premium (the level The Block reported in April 2026); the ETF charges 0.25% a year; the FCM requires initial margin of 35% of futures notional (an illustrative figure, since CME margins change often and the fund must use the live schedule); margin is posted in T-bills earning 4.0%.

  1. Size the spot leg. $50,000,000 / $86,000 = 581.4 BTC of exposure, bought as ETF shares.
  2. Size the futures leg. 581.4 / 5 = 116.3 contracts. The fund sells 116 standard contracts (580 BTC) and covers the 1.4 BTC residual with 14 micros or leaves it. Futures price: $86,000 x (1 + 0.05 x 90/365) = $87,060. Notional sold: 580 x $87,060 = $50.5 million.
  3. Gross carry. $87,060 minus $86,000 = $1,060 per coin; over 580 BTC, $614,800 locked in for 90 days if both legs are held to expiry.
  4. Costs. ETF fee: $50,000,000 x 0.25% x 90/365 = about $30,800. Commissions and exchange fees in and out: about $1,000. Net carry: roughly $583,000.
  5. Margin. 35% x $50.5 million = $17.7 million of T-bills, earning about $174,600 over 90 days. Total capital committed: $67.7 million.
  6. Annualise. $583,000 + $174,600 = $757,600 per quarter on $67.7 million, or 1.12% per quarter, about 4.5% annualised. That is roughly what a T-bill paid with no bitcoin, no FCM and no ETF. The excess return over cash is near zero, which is exactly why open interest fell through early 2026.
  7. Compare with November 2024. At a 20% basis the futures price would be $90,241, the basis $4,241 per coin and gross carry $2.46 million. Net of the same costs plus margin interest, the quarter yields about $2.6 million on $67.7 million, roughly 15.4% annualised. Same trade, same risks, more than three times the return.
Scenario Annualised basis Gross carry, 90 days Net return on $67.7m capital, annualised
April to October 2026 conditions 5% $614,800 About 4.5%
November 2024 conditions 20% $2,459,800 About 15.4%
Basis at zero at the roll 0% $0 on the new contract About 1.0% (margin interest minus ETF fee)
Basis at minus 2% at the roll -2% Minus $246,000 if rolled Negative; rational response is to unwind

Along the way. Suppose bitcoin rallies 20% to $103,200 in week three. The futures short loses 580 x $17,200 = $9.98 million, due in cash as variation margin within a day. The ETF has gained the same amount, but that gain is unrealised and sits at the prime broker, not the FCM. Without a liquidity line the fund must sell ETF shares to meet the call, forced out of a profitable trade at the worst moment.

When the basis turns negative. On a dated CME contract held to expiry the basis is locked; a mid-life move to backwardation is a mark-to-market swing, not a realised loss, because the short still converges to spot. The damage comes at the roll. If the next contract trades 2% annualised below spot, rolling means locking in a loss of about 0.5% for the quarter, around $246,000 here, while still paying the ETF fee and tying up $67.7 million. No rational carry trader rolls into that: they sell the ETF and buy back the future, and multiplied across the 37% of ETP assets hedge funds held in 2025 that is the sequence CF Benchmarks documented, compression then redemptions. In the perp version the equivalent is funding flipping negative, as on October 10, 2025; the short pays the long every eight hours until closed, and if ADL closes it first the hedge disappears while the spot leg remains.

How to evaluate a basis trade: a checklist

  • What is the basis net of the cash rate? A 5% basis against 4.5% T-bills is a 0.5% trade before costs. Compute the excess, not the headline.
  • Dated future or perp? A CME contract locks carry to expiry; a perp pays whatever funding turns out to be. October 2025 showed the difference is not academic.
  • Who holds the collateral? CME Clearing with T-bills, an offshore exchange holding stablecoins and a protocol vault with an ADL rule are three different risks. Price each.
  • Can the fund meet a 20% variation margin call in cash overnight? Spot gains are not at the FCM. Without a liquidity line the trade is a forced seller in a rally.
  • What is the roll plan? Decide in advance the basis at which the position is unwound rather than rolled, and what that does to reported ETF flows.
  • What does the COT say about crowding? Leveraged fund shorts above 60% of open interest, as on September 29, 2026, mean many participants will react to the same compression at once.
  • How are reference prices set under stress? Binance marked USDe collateral off its own thin order book on October 10, 2025 and paid $283 million for the error.

Risks and open questions

The basis trade is often called risk-free, and it is not. Direction is hedged; little else is. Margin risk is the most common failure: the short is marked to market daily in cash while the long’s gains are unrealised, so a sharp rally can force a liquidity-constrained fund out of a position that is making money on paper. Counterparty risk dominates offshore, where the exchange is clearing house, custodian and price oracle at once and where auto-deleveraging can close a hedge without consent, as on Hyperliquid on October 10, 2025. Funding risk applies to the perp version, where carry can go negative for weeks and a protocol like Ethena must draw on a reserve fund to keep paying.

The systemic questions are open. The trade links regulated ETF flows, CME positioning and offshore leverage into one loop, so stress travels between them within hours; the October 2025 event moved from Binance collateral marks to Hyperliquid ADL to ETF redemptions in weeks. Regulators have not said how they view hedged ETF holdings, and the SEC and CFTC split puts the two legs under different agencies with different data. How durable the mid-2026 rebuild is remains unclear: funding at 8.4% on October 6, 2026 against a 30-day average of 5.2% suggests quick re-levering, and each cycle so far has drawn more capital at a lower basis.

What to watch next

  • Weekly CFTC COT releases through Q4 2026. Leveraged fund net shorts growing alongside ETF inflows confirms carry capital re-entering; shorts shrinking into inflows would signal directional demand.
  • Q3 2026 13F filings, due by mid-November 2026. The first look at whether hedge funds rebuilt IBIT and FBTC positions during the Q3 rally or kept cutting.
  • CME December 2026 and March 2027 contract bases. A premium back above 8% to 10% would be expected to pull CME open interest off its April 2026 low.
  • Perp funding and Hyperliquid open interest in Q4 2026. Funding above 20% with record OI was the October 2025 setup; the September 2026 return to $14.3 billion makes the comparison live.
  • Federal Reserve decisions into 2027. Every cut lowers the hurdle rate and makes a given basis more attractive; hikes do the reverse, as the April 2026 compression showed.

Glossary

Basis
The difference between a futures price and the spot price of the underlying, usually quoted as an annualised percentage of spot.
Cash-and-carry
Buying the spot asset and selling futures against it to earn a positive basis with no net directional exposure.
Contango and backwardation
Contango is futures above spot (positive basis); backwardation is futures below spot (negative basis).
Funding rate
The periodic payment between longs and shorts in a perpetual swap that keeps its price near spot; positive funding means longs pay shorts.
Perpetual swap
A futures-like contract with no expiry, anchored to spot by funding rather than by convergence at maturity.
Commitments of Traders (COT)
The CFTC’s weekly report splitting futures open interest into trader categories such as Leveraged Funds and Asset Managers.
Leveraged Funds
The COT category covering hedge funds and CTAs; their short CME crypto positions are largely the futures leg of basis trades.
Variation margin
Daily cash settlement of futures gains and losses; a short must pay out in cash when the price rises.
Roll
Closing an expiring contract and opening the next; the moment a carry trade is re-priced at the new basis.
Auto-deleveraging (ADL)
An exchange mechanism that forcibly closes profitable positions to cover bankrupt ones when the insurance fund runs short.
CME CF Bitcoin Reference Rate (BRR)
The daily benchmark to which CME bitcoin futures cash-settle and which many spot ETFs use for NAV.

Why it matters

The basis trade is the hinge between the regulated ETF-and-CME world, where pensions and advisers now hold bitcoin, and the offshore perpetuals world where most leverage lives. Carry traders move capital between them, transmit stress from one to the other, set the premium at which leveraged longs are supplied, and generate much of the ETF flow data the market treats as a demand signal. Anyone reading inflows as conviction, or hedge fund shorts as bearishness, is reading the trade’s footprints without knowing what left them.

The record of 2024 to 2026 is a full cycle: a 25% basis that drew in tens of billions, a compression toward cash rates that pushed much of it out, a leverage event in October 2025, and a 2026 rebuild at thinner margins on more mature infrastructure. The useful question is not whether the trade is good or bad but how much of what you see in the data is really it. CCS covers the reporting and interviews behind this guide in the interview archive and on the research hub.

Sources

  1. CFTC Commitments of Traders via Tradingster: Bitcoin, CME (positions as of Sep 29, 2026), October 2, 2026
  2. CME Group OpenMarkets: Spot ETFs Give Rise to Crypto Basis Trading, November 5, 2025
  3. Crowdfund Insider: CME Group to Roll Out 24/7 Cryptocurrency Futures and Options Trading Starting May 29, May 2026
  4. CF Benchmarks: Revisiting the Bitcoin Basis, October 16, 2025
  5. CF Benchmarks: Tracking Bitcoin’s Flows, February 19, 2026
  6. The Block: CME Bitcoin futures activity slumps to 14-month low as basis trade unwind drains institutional demand, April 9, 2026
  7. The Block: Harvard leaves bitcoin ETF stake untouched in Q2, August 15, 2026
  8. The Block: Hyperliquid open interest climbs to $14.3 billion, September 8, 2026
  9. CoinDesk via Yahoo Finance: CME loses top spot to Binance in bitcoin futures open interest, December 22, 2025
  10. CoinDesk: Bitcoin is about to get a major bullish signal it hasn’t had in over a year, October 5, 2026
  11. FTI Consulting: The Crypto Crash of October 2025: When Leverage Met Liquidity, 2026
  12. The Defiant: The Ultimate 10/10 Crash Autopsy, October 2025
  13. 21Shares Research: Record crypto liquidations amid tariff shock, October 2025
  14. Decrypt: Binance Reimburses $283M After Market Crash and Asset Depegging Issues, October 13, 2025
  15. Blockworks: ‘Basis trade’ or HODLers: What’s behind the crypto ETF inflow spike?, April 28, 2025
  16. Hedgeweek: Brevan Howard ramps up bitcoin exposure amid institutional buying spree, August 2025
  17. sFOX: Bitcoin ETFs Face First Stress Test as Arbitrage Capital Exits, 2025
  18. DL News: ‘Basis traders’ are done dragging down Bitcoin price, says investment adviser, April 20, 2026
  19. Crypto Briefing: Bitcoin ETFs remain $1B short of breaking even in 2026, September 8, 2026
  20. LCX: Bitcoin ETF inflows hit $3.8B in strongest three-week stretch of 2026, September 5, 2026
  21. TFTC: Bitcoin futures leveraged funds short, CFTC data and the basis trade, September 14, 2026
  22. DefiRate: Bitcoin perpetual funding rate snapshot, October 6, 2026
  23. Crypto Coin Show: Ethena begins backing USDe with tokenized stocks via Binance basis trade, September 25, 2026
  24. Crypto Coin Show: CME’s 24/7 crypto launch will kill Bitcoin’s weekend gap, but Monday now matters more, May 27, 2026

Disclosure: This guide is for education only and is not investment, legal or tax advice.

Frequently asked questions

What is the crypto basis trade?

It is a cash-and-carry position: buy spot bitcoin, usually through a spot ETF, and sell an equal amount of futures. The trader has no net exposure to price and earns the basis, the premium of futures over spot. In early 2024 that premium approached 25% annualised on CME; by April 2026 it was near 5%, close to the Treasury bill rate.

Why are so many spot bitcoin ETF inflows described as hedged?

Because hedge funds buying ETF shares as the spot leg of a basis trade sell CME futures at the same time, so their net bitcoin exposure is zero even though the ETF records an inflow. Q1 2025 13F filings showed hedge funds holding about 37% of spot bitcoin ETP assets, and their IBIT holdings fell 28% in Q4 2025 as carry collapsed.

How can I see the basis trade in public data?

The CFTC's weekly Commitments of Traders report shows leveraged fund positions in CME bitcoin futures. As of September 29, 2026, leveraged funds were short 11,836 contracts and long 4,980, about 60% of open interest on the short side. Rising leveraged shorts alongside ETF inflows is the signature of carry capital entering.

What is the difference between basis and funding?

Basis is the premium on a dated futures contract, locked in when the trade is opened and realised at expiry when the contract converges to spot. Funding is the periodic payment on a perpetual swap, usually every eight hours, which varies continuously and can turn negative. On October 6, 2026 weighted bitcoin perp funding was about 8.4% annualised.

What happened to basis traders on October 10, 2025?

More than $19 billion of leveraged positions were liquidated in under 24 hours. Hyperliquid's open interest fell 56% and its auto-deleveraging closed profitable shorts, removing hedges. On Binance, USDe collateral was marked at about $0.65 for roughly 40 minutes, triggering liquidations; Binance later paid $283 million in compensation.

What happens to a basis trade when the basis goes negative?

On a dated CME contract held to expiry the original basis is locked in, so a mid-life move to backwardation is only a mark-to-market swing. The problem comes at the roll: rolling into a contract priced below spot locks in a loss, so traders unwind by selling the ETF and buying back futures, which appears as ETF outflows.

Is the basis trade risk-free?

No. Direction is hedged but margin, counterparty, funding and operational risks are not. A 20% rally on a $50 million position creates roughly $10 million of cash variation margin calls on the futures leg while ETF gains stay unrealised. Offshore, the exchange is clearing house, custodian and price source at once.

What returns does the trade produce in 2026?

At a 5% annualised basis, a $50 million position with 35% margin posted in T-bills returns roughly 4.5% annualised on total capital after ETF fees, about the same as cash. At the 20% basis seen in November 2024 the same structure returned about 15% annualised, which is why open interest rose then and fell through early 2026.

This explainer is reviewed and updated as the rules and the market change. Last reviewed October 6, 2026. It is educational content and not financial, legal or tax advice.

Keep learning