U.S. spot Bitcoin ETFs are recording simultaneous outflows and sovereign-linked accumulation, a divergence that Galaxy Research says reflects two very different investor playbooks rather than a single directional call on Bitcoin. Understanding which type of capital is leaving, and which is entering, changes how the headline flow numbers should be read.

Concentrated Selling Behind the ETF Outflows

According to Galaxy Research, institutions that file SEC Form 13F — the quarterly holdings disclosure required of institutional managers with at least $100 million in U.S. securities, filed within 45 days of quarter-end — held 355,000 BTC-equivalent in spot Bitcoin ETFs as of September 30, 2025. By June 2026, that figure had fallen to 307,000 BTC, a 14% decline over roughly nine months.

The selling was highly concentrated. Nine named firms — Millennium, Brevan Howard, Goldman Sachs, Capula, Farallon, D.E. Shaw, Schonfeld, Symmetry, and Tudor — together cut 68,000 BTC-equivalent. That figure alone exceeded the 48,000 BTC aggregate net decline across all roughly 2,200 reporting filers, meaning the rest of the filer base was adding exposure even as these nine exited.

Galaxy ties the fast-money exits to a specific mechanism: the economics of the cash-and-carry trade had deteriorated. The 30-day average annualized front-month CME basis — the spread between CME Group's near-dated Bitcoin futures and spot Bitcoin — fell from 7.9% before Bitcoin's all-time high to 4.8% in the 30 days through October 21, compressing the premium over three-month Treasury bills from 3.9 percentage points to just 0.6 percentage points. When carry barely clears risk-free rates, arbitrage-driven ETF positions lose their rationale.

Derivatives data reinforces the picture. CME asset managers' net long Bitcoin futures position fell from 43,300 BTC at the all-time high to 18,400 BTC, a 57% decline, while leveraged funds' net short unwound from 71,400 BTC to 30,600 BTC. Taken together, the moves suggest the entire basis-trade structure — long spot ETF paired with short CME futures — was being dismantled. An earlier Bitcoin ETF inflow streak had been built partly on that same arbitrage dynamic.

Sovereign Buyers Absorb the Supply

Galaxy names Morgan Stanley's wealth arm, Mubadala, Abu Dhabi's state-owned investment company, and the Abu Dhabi Investment Council, another investment arm of the Government of Abu Dhabi, among the buyers absorbing ETF shares as the fast-money crowd exits. These are not arbitrage desks hunting a 4.8% annualized return; they are long-duration allocators whose investment horizons are measured in years, not quarters.

The distinction matters for interpreting the net flow data. The aggregate 13F position fell by 48,000 BTC, but that headline number masks an underlying rotation: nine concentrated sellers shed 68,000 BTC while the broader filer base, including sovereign-linked names, added roughly 20,000 BTC in aggregate.

Sovereign and quasi-sovereign funds operate under disclosure constraints that limit what can be confirmed from public filings alone. Galaxy's analysis relies on SEC Form 13F submissions, which cover reporting institutions but exclude many advisers, family offices, and non-U.S. entities. The buyer list is therefore a partial picture, and the actual scale of sovereign accumulation may differ from what the filings show.

What the filings do show, according to Galaxy, is that non-filer Bitcoin holdings — a category that includes wallets and custodied positions outside the ETF wrapper — fell only 6.6% over the same period in which Bitcoin's price dropped 49%. That relative resilience suggests a class of holders with conviction strong enough to absorb a near-halving in price without proportional liquidation, behavior more consistent with strategic allocation than with momentum trading. Bears would note that falling prices could equally reflect holders who are unable or unwilling to realize losses, not necessarily strategic patience.

Can ETF Selling and Sovereign Buying Coexist?

The apparent contradiction — outflows alongside accumulation — dissolves once the ETF wrapper is understood as a neutral vehicle used by investors with opposite time horizons. That wrapper has been available as a regulated structure since the SEC authorized the first U.S. spot Bitcoin ETFs in January 2024, and it places no constraint on whether its holder is closing an arbitrage or building a reserve. An arbitrageur selling ETF shares to lock in a compressed basis return and a sovereign fund buying those same shares as a long-term reserve allocation are transacting in the same instrument for entirely different reasons. The net flow is negative; the composition of remaining holders is improving in terms of duration, though that qualitative assessment is the bull case and is not directly measurable from filings alone.

Bitcoin traded at $81,398 at the time of writing, up roughly 0.9% over the prior 24 hours, with a market capitalization of approximately $1.64 trillion and 24-hour trading volume near $25.5 billion, according to CoinGecko data. The Fear and Greed Index registered 51, squarely in neutral territory — consistent with a market that has absorbed the outflow headlines without triggering a panic response but has not yet re-rated on the sovereign buying narrative.

The Bull and Bear Cases

The bear case is straightforward. Aggregate 13F ETF holdings are still down 14%, the nine departing firms represented significant assets under management, and a compressed basis trade unwind can become self-reinforcing if more arbitrageurs reassess the same math simultaneously. A further narrowing of the CME basis toward zero would eliminate the remaining carry incentive and could trigger additional institutional redemptions before sovereign accumulation reaches a scale sufficient to offset the outflows.

The bull case rests on holder composition rather than flow direction. If the rotation from arbitrage capital to long-duration allocators is genuine, the remaining ETF ownership base is less likely to exit on price weakness alone, reducing future sell pressure for a given Bitcoin price decline. Galaxy's observation that non-filer holdings fell only 6.6% against a 49% price drop is the clearest data point supporting that thesis, though it remains an inference rather than a confirmed portfolio policy from any named sovereign entity. Separately, Bitcoin's recent consolidation above the $78,000 level amid hawkish Federal Reserve commentary suggests the price floor has so far been defended, but macro headwinds have not fully cleared.

What to Watch Next

Three data points will test the thesis: subsequent 13F filings covering the quarter ending September 2026; any official disclosure from Mubadala or the Abu Dhabi Investment Council regarding their Bitcoin ETF positions; and the CME basis spread relative to three-month T-bill yields. If the basis holds above risk-free rates, the incentive for further carry-trade unwinding diminishes; if it compresses further, the outflow cycle may not be complete.

ETF flows, in other words, are only one demand signal, and the composition of buyers and sellers inside the flow matters more than the net number. Sovereign exposure must be independently verified, given that Galaxy's analysis relies on Form 13F filings that cover reporting institutions but not all market participants. Timing and scale will determine market impact: the pace of sovereign accumulation relative to any further fast-money exits will decide whether the improvement in holder composition translates into price support.

Investors tracking the legislative dimension can also follow proposals such as the New Hampshire Bitcoin bond initiative, which illustrates how sovereign and quasi-sovereign Bitcoin exposure is being debated across multiple jurisdictions simultaneously.

Disclaimer: This article is for informational purposes only and does not constitute financial or investment advice. Cryptocurrency and digital asset markets carry significant risk. Always do your own research before making decisions.