The only comparable stretch on record ran from August 2022 into January 2023 and ended at the cycle low. Futures volume in July was just over $880 million against a February peak of $1.47 trillion.
Bitcoin's three-month futures basis has yielded less than a two-year U.S. Treasury note for 157 consecutive days, an inversion Glassnode says has only one precedent on record. That earlier stretch ran from August 2022 into January 2023 and ended at the cycle low.
The basis trade sits at the centre of institutional crypto. Sell a futures contract, buy the equivalent spot exposure through an ETF, and collect the annualised gap between the two prices with no directional risk. During the 2021 bull market that gap regularly paid twenty per cent or more. On Sunday it was worth roughly three per cent, against an average yield of 3.8 per cent on two-year Treasuries.
A dollar deployed into the trade now earns less than a dollar sitting in government paper. That is the entire investment case, and it is why the carry desks that ran the strategy for the last four years have been quietly closing books. Futures volume tells the same story: Coinglass shows July at just over $880 million, down from a February peak of $1.47 trillion. The mechanics are simple. If you can borrow at four and lend at three, you do not put the trade on. If you can earn 3.8 per cent on a Treasury note held to maturity, you do not tie up capital and margin at exchange counterparties for a rolling synthetic exposure paying three. The math is the same in every credit market; crypto has just discovered it applies here too.
The lower basis is not a broken market. It reflects one that has matured. When spot ETFs launched in January 2024, they gave anyone with a brokerage account a clean, cheap way to arbitrage the futures curve. Once tens of billions of dollars can hedge instantly, the mispricing that generated the 20 per cent carry compresses toward the cost of capital. What is left is a market that behaves like every other liquid futures market: narrow spreads, faster fills, fewer free lunches.
That maturation has a cost. The yield-chasing capital that inflated crypto derivatives during the last cycle is no longer being paid to stay. Basis trades are cash-neutral by design, but they are not fee-neutral for exchanges. CME, Deribit, Binance and Bybit built out their institutional products on the assumption that carry desks would keep rolling positions every quarter. Roll volume dries up when the roll costs money.
The composition of the losers matters. Basis desks at Jane Street, Jump, Cumberland and their crypto-native equivalents built teams and infrastructure explicitly to harvest this trade. When yields compressed in previous cycles, the same desks pivoted to funding-rate arbitrage on perpetuals or offered leveraged exposure to the funding trade to yield-hungry allocators. Both alternatives are structurally lower-yielding than the classic basis play, and both add credit and liquidity risk that Treasury notes do not.
The precedent Glassnode raised has substance. Between August 2022 and January 2023, bitcoin's futures basis stayed below two-year Treasuries for the same reason it is below them now: nobody wanted the leveraged long side of the trade. That period included the FTX collapse in November 2022, the price low of $15,800, and the beginning of the ETF-fuelled recovery. Correlation is not causation, but the composition of the market that produced that outcome and the market producing today's inversion is not obviously different: spot demand is exhausted, leverage has left, and the volumes reflect it.
Bitcoin traded just under $63,000 on Monday, down from July highs above $65,000. Spot ETFs bled a record $4.5 billion in June, and July is on pace for the weakest monthly inflows since launch. Both figures speak to the same withdrawal of directional buying that flattened the futures curve in the first place.
If the Federal Reserve delivers the September rate cut markets are pricing in, short-end Treasury yields fall and the basis comparison flips: a three per cent futures carry starts to look competitive against a two-year note in the low threes. Without a cut, the basis stays inverted and volumes keep contracting. Either way, the market that produced twenty per cent carry through 2021 is not coming back. The arbitrage window closes as an asset class becomes efficient, and the closing shows up in derivatives yield curves before it shows up anywhere else.
