In April we published The Dam and the River, an attempt to explain why Bitcoin had spent ten months drifting lower despite uninterrupted ETF inflows, growing corporate treasury demand, and a halving cycle that should mechanically tighten supply. The argument was structural rather than directional. When the SEC switched Bitcoin ETFs from cash creation to in-kind creation in July 2025, it handed Authorized Participants total control over the timing and sourcing of Bitcoin purchases. With a deep CME futures market available as a hedging tool, APs could absorb ETF demand quietly, satisfy their delivery obligations from private inventory and OTC desks, and never place a buy order on the public exchanges that actually set price.
The piece named one firm explicitly: Jane Street. Not because Jane Street had done anything wrong, but because Jane Street is the largest and most visible Authorized Participant in the Bitcoin ETF complex, and the specific market making behavior described in the article is precisely what its business is built to do. The conclusion was that the suppression mechanism was structurally temporary. OTC supply was draining. Mining could not keep pace. And the basis trade that connected the visible futures market to the invisible spot market had to terminate at some point.
On May 15, that point became visible in 13F filings. Jeff Park, CIO of ProCap and advisor to Bitwise, summarized the disclosure in seven words: "Price discovery is back on the menu."
Jane Street cut its IBIT (BlackRock's spot bitcoin ETF) position by approximately 71% during the first quarter of 2026. Its FBTC (Fidelity's) position fell by approximately 60% over the same period. These are not portfolio rebalancing moves at the margins. These are the largest and most consequential AP positions in the Bitcoin ETF complex being deliberately and quietly dismantled over the span of three months.
What a 13F Actually Shows
The 13F filing is a quarterly disclosure that institutional managers with more than $100 million in assets must file with the SEC within forty five days of quarter end. It captures long equity positions, including ETF shares, as of the last day of the quarter. It does not capture short positions, futures contracts, options, or any other hedging instruments held against those longs.
This matters enormously, because the IBIT and FBTC positions that appeared on Jane Street's previous 13F filings were almost certainly not directional bets on Bitcoin. A market maker the size of Jane Street does not take unhedged billion dollar exposure to any single asset. The long ETF position was one leg of a market neutral arbitrage (a paired trade built so bitcoin can rise or fall without changing the firm's profit: the money is made on the gap between the two halves, not on the direction). The other leg, invisible to the 13F, was a short position in CME Bitcoin futures of approximately equivalent notional size (the same dollar amount of bitcoin, expressed in contracts instead of coins).
This trade is sometimes called the basis trade, or cash and carry. The mechanics are unglamorous and extremely profitable when conditions allow. The AP buys the spot ETF, which represents real Bitcoin sitting in a vault. The AP simultaneously sells a futures contract that obligates them to deliver Bitcoin at a higher price several weeks or months later. The difference between the spot price and the futures price, divided by the time to expiry, is the basis. When the basis is positive and persistent, the trade prints money. The AP collects the spread at settlement. The position is delta neutral throughout: whichever way price moves, gains on one half offset losses on the other. The only real risk is funding cost (the cost of borrowing the money) and counterparty exposure (the chance the party on the other side fails to pay).
For a market maker, this trade has another important property. It absorbs ETF demand without requiring the AP to buy spot Bitcoin on the open market. When an investor buys IBIT, the AP can satisfy the in-kind creation obligation by sourcing Bitcoin from inventory, OTC, or a peer, while remaining economically hedged against the futures contract on the other side. The buying pressure that would have appeared on exchanges under the old cash creation model gets distributed across private channels and the futures market. The spot price stays quiet.
What a 71% Cut Means in Practice
When an AP unwinds a basis trade, both legs close together. The long ETF position is sold. The short futures position is bought back. The arbitrage is dissolved. If only one leg closes, the AP is suddenly running directional exposure, which violates the entire premise of market making.
It is tempting to read the cut as a basis-trade unwind: if the long ETF leg shrinks by 71%, the offsetting short futures leg has to shrink with it, and the arbitrage pathway that bypassed spot closes. That is the reading this article leans toward. But it is an inference, not an observation, and honesty requires naming what the filing cannot show. A 13F reports only the long equity leg. The short futures position, the thing that would actually confirm a basis-trade unwind, is invisible to it. Several other explanations fit the same disclosed numbers:
- The long leg moved, it did not close. Jane Street operates through many legal entities. The position could sit on a filer that this 13F does not consolidate, or have shifted to a book that does not report here at all.
- An AP or issuer rotation. Another Authorized Participant may have taken over the creation and redemption role, in which case the basis trade simply continues under a different name.
- A hedging-venue shift. The hedge may have moved from CME futures to options or to offshore perpetual swaps (futures contracts that never expire, traded on venues outside US regulation), shrinking the visible ETF footprint without ending the strategy.
- An ordinary risk decision. If part of the position was not fully hedged, the cut could be plain risk reduction with nothing to do with the plumbing.
None of these can be ruled out from a long-only quarterly snapshot. The basis-trade-unwind reading is the one most consistent with the rest of the picture: the draining OTC desks, the compressing basis, the timing. It is the working thesis here, not a proven fact.
Why? The tweet does not say, and the 13F does not say, and Jane Street will not say. The reasons matter less than the consequence, but several are plausible: the basis spread has compressed as funding rates fell and competition for the trade rose, financing costs have risen, OTC supply has grown unreliable enough that the AP cannot confidently source Bitcoin at its quoted creation prices, regulatory attention is intensifying, and capital may be moving elsewhere. Any one would suffice. Several may be true at once.
What matters is the structural consequence. With the largest AP scaling back the basis trade, future ETF inflows lose access to the most efficient hedging mechanism that has absorbed them for the past year. The next wave of demand must be intermediated through a thinner, more expensive, more fragile market structure. And the only path that always remains open is the one APs have spent the entire in-kind era avoiding: buying Bitcoin on the public spot market.
The Convergence Was Always Going to Arrive
The original article argued that three forces were converging that would eventually overwhelm the dam. The OTC pool was draining, mining could not keep pace, and the basis trade had to stay profitable to keep absorbing flows. Six weeks later, the third force has materialized as a fourth, visible fact: the Authorized Participants that built the dam are themselves stepping back.
OTC supply continues to deteriorate. CryptoQuant's on-chain tracking now shows OTC desk balances at roughly 115,000 Bitcoin, down from approximately 156,000 in July 2025 and from a 486,000 peak in September 2021. At the recent average outflow of roughly 276 Bitcoin per day, a naive straight-line projection empties the remaining pool within about a year. That projection is a rough horizon, not a date. A draining flow has no reason to stay linear, desks can be replenished from miners or other sellers, and OTC is only one of several sourcing channels. What holds up is the direction; the timing is not. What is fair to say is that the buffer APs relied on to source Bitcoin without touching public markets is much thinner than it was. Any incremental ETF demand sourced through OTC is bidding against Strategy, the ETFs themselves, and a growing list of sovereign and corporate treasuries.
Halving math has not changed, but institutional demand has accelerated. Roughly 450 Bitcoin per day enter the system through mining, about 164,000 per year. ETFs collectively hold over 1.27 million. Strategy has continued to compound through 2026, rising from roughly 673,000 Bitcoin at the start of the year to approximately 843,000 by mid May. That is more than 170,000 Bitcoin acquired by a single corporate buyer in less than five months, exceeding an entire year of new mining issuance, financed by what is now the largest US equity raise of 2026. The annual new supply is no longer a small fraction of institutional demand. For stretches of this year it has been smaller than one company's quarterly purchases.
The composition of that demand has also shifted in a way that reveals exactly what the suppression mechanism did and did not accomplish. In April 2026, Strategy quietly passed BlackRock's IBIT to become the largest single Bitcoin holder in the world, with approximately 815,000 coins to IBIT's 803,000 at the time. During the same six month window in which IBIT's holdings barely moved and Bitcoin's price fell more than fifty percent, Strategy added roughly 80,000 Bitcoin at depressed cycle prices. The suppression mechanism did not suppress aggregate demand. It redirected it. Passive ETF buyers got front run by an active accumulator that understood exactly what the dam was doing, and was happy to buy quietly while the public market was unable to find a clearing price.
The basis is closing, and APs are leaving with it. When the futures premium compresses, the cash and carry trade stops being profitable. When the cash and carry trade stops being profitable, the AP no longer has a reason to hold the long ETF leg. When the AP unwinds the long ETF leg, the entire arbitrage pathway that absorbed ETF flows quietly dissolves. The mechanism that allowed Bitcoin to trade as if its demand and supply curves had decoupled cannot survive its own operators losing interest in operating it.
What "Price Discovery Is Back on the Menu" Actually Means
Park's phrase is unusually precise for social media. He did not say price is going up. He did not say Bitcoin is undervalued. He said price discovery, the underlying market function, is returning to a venue where it had been temporarily disabled.
Price discovery is the process by which a market aggregates buyer and seller intent into a single quoted price. It requires that the largest sources of demand and supply both express themselves in the same venue. For the last ten months, the largest source of incremental Bitcoin demand, namely ETF inflows, has been intermediated through a structure that allowed it to express itself almost entirely in private channels. Spot prices were therefore being set by a residual market, not the marginal demand curve. In plain terms: the printed price came from whoever happened to still be trading on the public exchanges, while the biggest buyer did its shopping in a back room.
When the basis trade unwinds and APs source Bitcoin through the only channel still capable of supplying institutional size, namely public exchanges, the marginal demand and supply curves meet again in the venue where price is quoted. The result is not necessarily a rally. The result is a price that reflects what investors and holders actually believe Bitcoin is worth, rather than the residual of a hedging arrangement.
Given everything visible about the supply side, the halving schedule, OTC depletion, corporate accumulation, and the broader macro context, the most reasonable expectation is that this restored price discovery process will resolve upward. But the certainty in that sentence belongs to the words "price discovery", not to the word "upward". The market's job is to find a clearing price. For ten months it has been prevented from doing so. That prevention is now ending.
The Timing Just Changed
The original article closed by observing that the supply was draining, the demand kept building, and the only question had ever been timing. The 13F filing did not change the supply situation, the demand situation, or the mathematics of the halving. It changed the timing.
The dam was already full. The reservoir behind it was already rising. The open question in April was when the operators would step away from the controls. They began stepping away three months ago. The Q2 2026 13F, due in mid-August, will show whether the trend has continued, reversed, or accelerated. None of those outcomes leave the suppression mechanism intact in its previous form.
Price discovery is back on the menu. On the reading argued here, the dam looks closer to breaking than to holding. That is a thesis, not a readout, and the next 13F is where it gets tested rather than asserted.
Tracked: the Q2 13F call and the OTC-depletion call from this piece are logged on the predictions scorecard with dated resolution conditions. Hold them to it.