Who Buys Bitcoin After Strategy

The memo and the sale

On July 1, 2026, Bitwise chief investment officer Matt Hougan sent a memo that traveled fast. For years, he wrote, Strategy had been the most dominant Bitcoin buyer in the world, a steady one-way source of demand. Those days, he added, are likely over.

A week later, Strategy did something it had done only rarely in its history. It sold Bitcoin. The company disposed of 3,588 BTC for roughly $216 million to help fund preferred dividends, its largest sale to date. Bitcoin did not fall on the news. It rose.

That reaction is the real story, and it is not the one most headlines chased. The question worth asking is not whether Strategy is failing. By most measures it is not. The question is what changes for Bitcoin when the marginal buyer changes.

This is an analysis of that shift, not a prediction. CanoeBit takes no view on where the price goes next. What follows is about mechanics: who has bought Bitcoin, how each buyer's demand actually worked, and why the handover now under way looks structurally different from the ones before it.

A relay of buyers

Hougan frames Bitcoin's history as a relay. The network's properties have never changed. The dominant buyer has, again and again. His sequence runs from the cypherpunks to Asian retail, then to US retail, then to the Grayscale Bitcoin Trust, then to Strategy, and now, he argues, to institutional allocators.

Each cohort was once met with the same skeptical question: who is left to buy? Each time, a larger pool of capital answered. This relay is the thread of an earlier piece, Bitcoin Isn't the Money You Were Promised, which traced how ownership keeps moving outward.

The useful part is not the list. It is that the three most recent cohorts each demanded Bitcoin through a completely different mechanism. Those mechanisms are what make the current handover legible.

The trust that became a one-way seller

Grayscale's GBTC was, for years, a trap that happened to work in Bitcoin's favor. As a closed-end trust, it let investors buy in but offered no way to redeem shares for the underlying Bitcoin. Capital could enter. It could not easily leave. At its peak the trust held around 617,000 BTC when it converted to an ETF in January 2024.

Because there was no redemption, GBTC often traded at a steep discount to the Bitcoin it held, wider than 40 percent during the 2022 and 2023 downturn. The structure locked holders in and kept their Bitcoin off the market.

Conversion reversed the mechanism. Once redemption became possible, that same trapped capital rushed for the exit through redemption and arbitrage. GBTC has since shed more than 60 percent of its Bitcoin, with cumulative outflows near $25.9 billion, worsened by its category-high 1.5 percent fee. The trust that had been a one-way holder became a one-way seller. The reflexivity did not disappear. It flipped direction.

The company that became a one-way buyer

Strategy ran the opposite engine, and for years its fuel was the premium the market placed on its own shares.

For years the market valued Strategy above the Bitcoin on its balance sheet, a premium captured by a multiple known as mNAV. That premium was the machine. Strategy could sell new common stock through at-the-market programs, or issue low-coupon convertible notes, collect more dollars than the Bitcoin those securities represented, and use the difference to buy still more Bitcoin. Each turn added Bitcoin per share. At-the-market common-stock sales were the primary tool for years, alongside roughly $8.2 billion in near-zero-coupon convertible notes, part of more than $46 billion in external capital the company has raised to buy Bitcoin since 2020.

In 2025 Strategy layered on a newer instrument: perpetual preferred stock, ending with STRC, a variable-rate series built to trade near $100 while paying a high dividend. By early 2026 STRC had become its main funding vehicle, shifting the engine from share dilution toward preferred issuance. That shift mattered, because it moved a growing share of the funding onto a security with a fixed dividend to defend.

The loop only runs in one direction, and only above one line. When the premium compresses toward 1.0x, issuing common stock stops adding value and starts diluting existing holders. The flywheel stalls. And a preferred stock that slips below its target price makes new issuance there difficult too.

That is what June exposed, on both funding channels at once. STRC broke sharply from its $100 target and touched an intraday low near $71.25 on June 26, 2026. Bitcoin fell to $58,190 on June 25, its lowest level in 21 months. On June 29 Strategy unveiled its Digital Credit Capital Framework: a $2.55 billion dollar reserve, buyback authorizations, and permission to monetize up to $1.25 billion of Bitcoin to fund reserves, dividends, and repurchases. It also stopped automatically defending STRC's $100 level. The company that famously would not sell had built the machinery to sell.

Then it did, with the $216 million disposal in early July. The acute stress has since eased. By July 15, 2026, mNAV had recovered to about 1.02, STRC traded back near $90, and MSTR had closed around $97.58 the day before, still down roughly 36 percent on the year. These are snapshots of a fast-moving position, not fixed facts.

The mechanism to hold onto is this: Strategy's demand was a reflexive, financed, one-way bid. It works while the premium holds and it inverts when the premium does not.

The allocators who buy differently

Institutional demand does not work like either of those. It is not a levered flywheel and it is not trapped capital. It is mandate-driven: slower to arrive, more diffuse, and far stickier once it does.

The early evidence is real but still small in proportion. Morgan Stanley has launched its own Bitcoin ETFs. Wells Fargo has added Bitcoin to model portfolios. Texas became the first US state to fund a Bitcoin reserve. The US Strategic Bitcoin Reserve holds roughly 328,372 BTC, the largest known state holding, assembled from seizures, with an executive order directing budget-neutral additions. Abu Dhabi's Mubadala disclosed about $565.6 million in IBIT shares as of March 31, 2026. BlackRock guides conservative clients toward a 1 to 2 percent allocation. Measured against gold, still the incumbent benchmark these allocators reach for, Bitcoin remains a small position, a comparison explored in Bitcoin vs. Gold.

This cohort is not a smooth one-way bid either. BlackRock's IBIT peaked above 806,700 BTC in April 2026 and fell to around 729,956 by early July, and the ETF complex just went through its worst outflow streak on record before a tentative July reversal that IBIT led. One concentration is worth naming plainly: Coinbase Custody holds the IBIT Bitcoin.

But the difference in kind matters more than the volatility. Allocator demand is spread across many mandates rather than staked on one balance sheet, and it does not depend on a premium staying above a line. It is allocative, not reflexive.

What actually changes

Set the three mechanisms side by side and the shift becomes clear. Bitcoin is moving from demand that was concentrated and reflexive toward demand that is broad and allocative.

Concentrated, reflexive demand is explosive on the way up and fragile at the turn, because it is levered to a single premium. Broad, allocative demand is slower and less dramatic, but it does not carry the same reflexive trigger. That is the structural change. It is not a statement about price, and CanoeBit is not making one.

The honest other readings

There are serious arguments that this shift is being overread, and they deserve equal weight.

Strive's Matt Cole argues the STRC episode drew outsized attention, noting Strategy's roughly 847,363 BTC is about 4 percent of supply, below the 5 percent threshold traditional finance treats as notable ownership. HashKey's Tim Sun and Grayscale's Zach Pandl both argue a slower Strategy is healthier, because it unwinds a financing-driven distortion and lets a firmer floor form from real demand. Bernstein sees a forced sale as unlikely, with debt at roughly 13 percent of Bitcoin holdings and no principal due until the third quarter of 2028. Standard Chartered's Geoffrey Kendrick reads the whole affair as a communication problem rather than a balance-sheet one.

Hougan himself is not bearish. He expects Strategy to stay a net buyer over time and reads the July sale as a rational actor lowering the odds of a forced liquidation, which is part of why Bitcoin rose on the news.

The counterarguments are real too. Peter Schiff contends that selling into weakness can multiply losses and that the business model has fundamentally changed. JPMorgan warns that formally allowing sales turns Strategy into both buyer and seller, adding avoidable two-way risk. Estimates of where genuine balance-sheet stress would begin also diverge widely: Hougan put it near a 70 percent fall to roughly $18,500, while Strategy's chief executive, Phong Le, pointed to the $8,000 to $10,000 range on July 14, closer to an 85 percent drop. Both are estimates, not thresholds anyone has tested.

Where this reading breaks

An honest analysis names the conditions that would prove it wrong.

This reading would not hold if Strategy resumes aggressive accumulation as prices fall, and the mNAV compression turns out to have been a brief episode rather than a structural ceiling. It would also fail if institutional allocation stays marginal for years. In that case the baton has not truly passed, and calling this a cohort change was premature.

For now, the mechanism is what is visible. One dominant, financed, reflexive buyer is stepping back, and a wider field of allocators is stepping in. How much they buy, and how fast, is the open question. It is not one this analysis will pretend to answer.