Mentioned in This Article
1. Who sold Bitcoin ETF shares during the 2026 slump?
A Galaxy Research report published on Monday, October 26, 2026 finds that nine hedge funds and banks, led by names such as Millennium Management, Goldman Sachs and Tudor Investment, cut 68,000 bitcoin (BTC) of spot ETF exposure, more than the entire 48,000 BTC net decline among all 13F filers. Morgan Stanley's wealth arm and sovereign funds were among the buyers.
The finding cuts against a common reading of the slump, in which broad institutional selling is blamed for the drop in bitcoin's price. According to the report, institutional selling of U.S. spot bitcoin exchange-traded funds was concentrated in a small group of firms, while wealth platforms and sovereign investors were net buyers of the same products.
Galaxy's data comes from 13F filings, the quarterly disclosures that large U.S. investment managers file to report their equity holdings, including ETF shares. Those filings run only through June 2026, so they do not capture the most recent shifts in positioning, and Galaxy noted that it holds bitcoin itself.
Spot bitcoin ETFs are exchange-traded funds that hold actual bitcoin and trade like stocks, which lets institutions and individuals gain exposure through a brokerage account. Because their holdings are reported in shares, the quarterly 13F filings offer one of the few public windows into who owns them, and Galaxy used that window to separate the holders by type.
2. How much bitcoin did institutions hold, and how much did they cut?
At the end of September 2025, just before bitcoin reached its all-time high, 13F filers collectively held 355,000 BTC in spot ETF shares. By June 2026 that total had fallen to 307,000 BTC, a decline of 14%.
Nine firms account for more than all of that reduction: Millennium Management, Brevan Howard, Goldman Sachs, Capula Management, Farallon Capital, D.E. Shaw, Schonfeld Strategic Advisors, Symmetry Investments and Tudor Investment. Together they sold 68,000 BTC, worth roughly $5.4 billion at current prices, and trimmed their positions by between 29% and 100%.
The rest of the filer universe moved the other way. About 2,200 other 13F filers added roughly 20,000 BTC, worth about $1.6 billion. That is why the net decline across everyone, 48,000 BTC, is smaller than the 68,000 BTC the nine firms removed.
The distinction between the nine sellers and everyone else is the heart of the report. A headline decline of 48,000 BTC across all filers could suggest a broad retreat, but the underlying numbers show a small group moving decisively in one direction while a far larger group moved modestly in the other. The 29% to 100% range shows that some of the nine firms trimmed their positions while others exited entirely.
3. Who stepped in as buyers?
Morgan Stanley's wealth management arm, Mubadala and the Abu Dhabi Investment Council were named among the buyers of spot bitcoin ETF shares. Mubadala and the Abu Dhabi Investment Council are sovereign investment vehicles of Abu Dhabi, and Morgan Stanley's wealth business serves private clients through its advisers.
Holders closer to the retail end of the market were notably steady. Financial advisers, family offices and individuals who do not file 13F reports cut their holdings by only 6.6%, even though the price of bitcoin fell 49% over the period. In other words, the investors least likely to be forced out by a fast-moving trade held on through a drawdown of nearly half.
Sovereign wealth funds are state-owned investment pools that typically hold assets for the long term, and wealth management platforms allocate on behalf of individual clients. Both tend to behave differently from hedge funds, which can enter and leave positions quickly when a strategy stops paying. The presence of such buyers during a slump points to a shift in who holds the funds.
4. Why did the hedge funds leave? The cash-and-carry trade explained
Alex Thorn, Galaxy's head of firmwide research, argues that the sellers were mostly running a trade rather than expressing a view that bitcoin would fall. Many of them used a strategy known as cash-and-carry: buying spot ETF shares while shorting CME bitcoin futures, so that the position is hedged against price moves and earns the gap between the two, known as the basis.
That gap is the profit source, and it has narrowed sharply. In the 30 days before the all-time high, the front-month futures basis averaged 7.9% annualized, or 3.9 percentage points above three-month Treasury bills. In the 30 days through October 21, it averaged 4.8%, only 0.6 points above bills. Galaxy estimates that after fees, the cost of rolling contracts and margin requirements, the trade now earns about 1.5 points over cash.
Thorn's conclusion is that the arithmetic, not the price, drove the exits. In his words, "That arithmetic, not the price, is why the multi-strategy desks left." A position that offers little more than Treasury bills while tying up capital and carrying operational risk is unattractive to funds that can deploy their money elsewhere.
To see why the basis matters, consider how the trade works. A fund that buys ETF shares and shorts the same amount of bitcoin futures is protected from price swings, because a loss on one side is offset by a gain on the other. What remains is the difference between the futures price and the spot price, which converges as the contract nears expiry. When that difference is wide, the trade pays handsomely for a low-risk position. When it narrows to within a point of Treasury bill yields, the same capital could earn almost as much in government debt without exposure to exchange, margin or rolling costs.
5. What does futures positioning show?
Data from the Commodity Futures Trading Commission supports the story. Asset managers' net long position in CME bitcoin futures dropped 57%, from 43,300 BTC in the week of the all-time high to 18,400 BTC in the latest reading. Leveraged funds' net short fell by the same proportion, from 71,400 BTC to 30,600 BTC.
The two moves are what a shrinking carry trade would produce: as funds unwound the long ETF leg, they also closed the matching short futures leg. Open interest in CME bitcoin futures has fallen 28% in BTC terms and 51% in dollar terms, the larger dollar decline reflecting both fewer contracts and a lower bitcoin price.
Longer-dated contracts do not change the picture much. November and December futures priced at 6.3% to 6.5% at the October 21 close, roughly 2 points above bills before costs. Thorn argues that this thin edge does not compensate for tail risk, citing December 2022, when the basis inverted to negative 31% annualized.
The 2022 episode Thorn cites shows what the tail risk looks like. A negative 31% annualized basis means futures traded far below spot, a condition that can inflict losses on a carry position when it needs to be closed. Against that history, earning about 2 points over bills on December contracts leaves little reward for the risk of such a dislocation.
6. Did overall ETF holdings collapse?
They did not. Spot bitcoin ETF holdings fell only 11% from the October 2025 peak to the July 2026 trough, while the price of bitcoin dropped 53%. The gap between the two figures indicates that most capital stayed in the funds.
Since the June low, funds have added 87,000 BTC, which Galaxy characterizes as directional buying rather than carry. That is a different kind of demand from the basis trade: investors are taking outright exposure to bitcoin's price rather than harvesting a spread.
Recent daily flows have been uneven. Outflows of nearly $1 billion over two weeks were followed by a strong day on October 19, led by BlackRock's iShares Bitcoin Trust (IBIT), and then by roughly $1 billion in inflows over three days, which gave way to heavy outflows. Flows turned positive again this week, with about $286 million in inflows so far.
The contrast between an 11% decline in holdings and a 53% decline in price carries a simple message about investor behavior. Funds in aggregate held on through most of the drawdown, and the buyers who arrived after the June low added 87,000 BTC, a larger amount than the 68,000 BTC sold by the nine firms. Galaxy's reading is that new demand is outright and directional, unlike the hedged demand it replaced.
7. What does this mean for bitcoin's next move?
The report suggests that the headline picture of institutions fleeing bitcoin ETFs overstates the case. A narrow group of arbitrage-driven sellers accounts for the net decline, and their exit followed a drop in the profitability of their strategy rather than a change in their opinion of bitcoin.
Galaxy's data has limits. The 13F filings end in June, so they cannot show what the nine firms or the new buyers have done since. Future 13F filings for the third quarter will show whether the pattern held, and the basis, flows and CFTC positioning are the indicators to follow in the meantime.
For investors following ETF flows, the report is a reminder that a withdrawal from a fund does not always signal a negative outlook. When the seller is running a hedged spread, the sale can reflect nothing more than a change in the economics of the trade. Reading flows alongside futures positioning and the basis, as Galaxy does, gives a fuller picture of whether the money is leaving because of price expectations or because a strategy has simply stopped paying.