The concentration risk embedded in corporate balance sheets has converted from theoretical tail to live supply-side variable, with Strategy alone holding over 840,000 BTC as of July 2026 after disposing of approximately 3,620 BTC. For algorithmic execution desks, the structural input has inverted: corporate accumulation no longer functions as a price floor, it functions as an open sell queue.
The unwind, by the numbers
Exit velocity varies sharply across the issuer cohort, and the dispersion is itself the signal:
- Strategy (formerly MicroStrategy): Sold ~3,620 BTC, retains >840,000 BTC. Largest concentrated position in the corporate set and the primary overhang risk to spot BTC.
- Satsuma Technology: Full liquidation of 668 BTC approved, triggering delisting from the London Stock Exchange. Terminal event for the smallest cohort member.
- Sequans Communications: Sold 1,025 BTC plus roughly 80% of remaining holdings, proceeds directed at convertible debt repayment. Forced deleveraging rather than strategic rotation.
- MARA Holdings and Bitdeer: Net sellers of BTC, funds redeployed to AI data center buildout and share repurchases.
- Twenty One Capital: CEO Jack Mallers resigned. Governance signal preceding any further balance sheet action.
- Bitcoin Standard Treasury Company (BSTR): Merger proposal scrapped. Capital structure plan withdrawn before execution.
Mean reversion of treasury company valuations against NAV is no longer uniform. It is entity-specific and execution-quality dependent. Strategy retains optionality; Satsuma and BSTR have hit structural termination.
Why the AI pivot is mechanical, not narrative
Bitcoin mining operations already possess the physical layer AI workloads require: power infrastructure, cooling systems, and data center shells. Conversion from proof-of-work to GPU-dense training or inference is redeployment of fixed assets already on the balance sheet, not a rebranding exercise. The market, however, is pricing these pivots as forced rotations. The signal sits in the equity tape, not the press release. Whether miners attract an AI infrastructure multiple or retain a distressed-crypto multiple will be determined by realized revenue per megawatt, not by announcement cadence.
Risk-adjusted verdict
Three variables deserve continuous monitoring in any systematic BTC strategy recalibrated for this regime:
1. Overhang delta. Track Strategy's BTC balance week-over-week. Any reduction from the 840,000 baseline functions as incremental supply hitting the order book. Slippage on a single large liquidation event will be non-trivial against current 24-hour volumes.
2. Solvency-versus-strategy sequencing. Sequans' convertible debt repayment is the template that generalizes across the cohort: balance sheet survival precedes narrative pivot. Algorithms that price treasury equities on AI optionality before debt service is resolved will misprice the equity leg.
3. Correlation regime break. If treasury company equities decouple from BTC NAV and begin trading on AI-infrastructure multiples, the historical hedge coefficient degrades. Backtests anchored to BTC-equity correlation over the prior cycle will register slippage in the live conversion.
The data is unambiguous: corporate BTC treasuries have transitioned from accumulation engine to liquidation queue. Execution algorithms priced for the prior regime require parameter reset for the supply profile this implies.