Share
Galaxy Research says sovereign funds absorbed what hedge funds sold during Bitcoin's slump, but quarter-by-quarter filings from two of its named holders show both sides buying through December.
Key Takeaways
- Millennium and Mubadala both added IBIT in Q4.
- Their paths parted only in early 2026.
- Millennium’s stake fell about 72% from December.
- Mubadala’s share count was flat in Q2.
Galaxy set out the argument in its November 6 research brief, which draws on an October report comparing 13F filings at September 30, 2025, and June 30, 2026.
Galaxy’s endpoints have a logic. September 30, 2025, was the last quarter-end before Bitcoin’s October peak, and the cycle low came on June 30, 2026. Across that span, Galaxy says, ETF holdings reported by 13F filers fell 14%. It names Millennium Management among nine multi-strategy, macro and bank trading desks that cut 68,000 BTC of exposure between them, and Mubadala Investment Company among the wealth platforms and sovereign funds that bought.
On those two dates, the filings bear it out. Millennium’s ordinary shares in BlackRock’s iShares Bitcoin Trust (IBIT) fell about 53%, while Mubadala’s rose about 69%. The December and March filings in between change the order of events.
Sources: SEC Form 13F filings linked above; Mubadala’s December 2025 count as reported by CoinDesk from its 13F. Ordinary shares only. IBIT mark is the reported position value divided by shares held at each quarter-end, rounded to whole dollars.
In the December quarter, with IBIT’s quarter-end mark down by about a quarter, both managers bought. Millennium’s December count was the largest it had reported at any quarter-end since IBIT launched in January 2024.
The paths parted in the March quarter. Millennium sold 44% of its position while Mubadala kept adding, the only three months in this comparison when the two moved in opposite directions. In the June quarter, Millennium cut again and Mubadala held still: its March and June filings both show exactly 14,721,917 shares.
Measured from December rather than September, Millennium’s reduction is roughly 72%, not 53%. Galaxy’s endpoints understate how much the fund sold, and they leave out that it was buying alongside the sovereign investor three months earlier.
Mubadala’s stake grew while its value shrank
The IBIT mark explains a result that looks contradictory in Mubadala’s filings. Its position was reported at $567 million in September 2025 and $490 million in June 2026, about 14% less, even though it held 69% more shares. Each share was marked about 49% lower over that period, matching the 49% decline in Bitcoin that Galaxy cites.
Those marks are quarter-end prices, not purchase costs. The filings do not show what Mubadala paid for the roughly 6 million shares it added, and the positions are fund shares rather than bitcoin held directly, a distinction Coindoo examined in what owning a crypto ETF actually gives investors. For tracking whether a position grew, the share count is the cleaner measure.
The same arithmetic separates selling from price on Millennium’s side. Most of the drop in its reported IBIT value between December and March came from shares it no longer held, not from the lower mark.
Millennium’s IBIT value, December 31 to March 31
Calculated from Millennium’s December and March 13F filings; figures rounded.
A fund can trade Bitcoin ETFs without a view on Bitcoin
Building the largest stake it had reported into year-end, then unwinding most of it within six months, fits a trade that does not depend on Bitcoin’s direction. In a basis trade, outlined in CME Group’s explainer, a fund buys spot exposure such as ETF shares and shorts Bitcoin futures, collecting the gap between the two prices. When the gap narrows, the trade is closed by selling the ETF and buying back the futures.
Form 13F shows only the first half of that trade. The SEC’s 13F guidance excludes short positions, and Bitcoin futures, which trade under CFTC oversight, are not 13F securities. A basis unwind therefore appears in the filings as plain ETF selling. The IBIT options Millennium did disclose were small next to its shares: from September 2025 through March 2026, its call and put rows each covered fewer than 1.5 million underlying shares. Mubadala’s June filing lists a single IBIT row with no options attached.
Aggregate futures data covers part of what the filings omit. Citing the CFTC’s Traders in Financial Futures report, Galaxy says leveraged funds’ net short in CME Bitcoin futures, the futures leg of ETF basis trades, shrank from 71,400 BTC in the week of the October 2025 high to 30,600 BTC by late October 2026. Over the 30 days before the high, the front-month basis averaged 7.9% annualized, 3.9 points above three-month Treasury bills. In the 30 days to October 21, it averaged 4.8%, just 0.6 points above bills. Galaxy concludes that this shrinking margin is why multi-strategy desks left.
That is an industry-wide explanation, and Millennium’s timing is consistent with it. No filing ties Millennium’s sales to a futures short, though. The same limit applies on the buying side: Mubadala’s single share row shows a larger holding, not how long it plans to keep it.
What the September filings can settle
Third-quarter 13F reports are due November 16 and cover holdings on September 30, 2026. They are the first filings to show who held ETF shares after the June low.
Galaxy, which told clients in October that it believes Bitcoin’s bottom is in, reads the 87,000 BTC the ETFs have added since that low as directional buying rather than carry, because leveraged funds’ futures shorts fell over the same stretch. The new filings put names to part of that demand.
Two counts from this comparison are the first to check.
Reporter at Coindoo
Alex is Editor-in-Chief of Coindoo and co-founder of Millennial Media Group, with nearly a decade of experience covering financial markets - crypto first, then everything else. It started in 2016 with Bitcoin. Like most people at the time, he didn't fully understand it - so he kept digging. Blockchain, tokenomics, the projects, the cycles. That curiosity never stopped, and eventually pulled him into traditional markets too: equities, commodities, macro. Not because he left crypto behind, but because you can't properly understand one without the other. What drives him is straightforward: he wants to know why something is happening, not just that it's happening. Most market coverage stops at the headline - price up, price down, here's a chart. Alex finds that kind of reporting actively unhelpful. If you walk away from an article without understanding the mechanism behind the move, what did you actually learn? He holds a degree in Tourism from New Bulgarian University - not the most obvious path into financial markets, but markets have a way of pulling in people who are simply too curious to stay out. He has authored over 200 in-depth analyses and more than 10,000 articles across crypto and traditional finance. He still thinks every day in markets teaches him something new. That's probably why he hasn't stopped.
Related stories
Advertisement
Up to 11100 usdt bonus for new users.
Low trading fee.
