Michael Saylor sold Bitcoin. He said he never would. The data says otherwise.
Strategy (formerly MicroStrategy) held 840,447 BTC as of August 16, 2025. Their average purchase price: $75,385. Total cost: $63.36 billion. Then came the Q2 net loss of $8.22 billion. Stock down 40% year-to-date. And then the sale. The man who built a personal brand around "never sell" disposed of some of the stack. The narrative is cracking.
s heart.
Context: The Leveraged Bitcoin Vehicle
Strategy is not a Bitcoin miner. It is not a protocol. It is a publicly traded company that buys Bitcoin using debt and equity dilution. Since 2020, Saylor has positioned the company as a proxy for Bitcoin exposure, but with a twist: leverage. By issuing convertible bonds at low interest rates and selling stock via ATM programs, he accumulated the largest corporate Bitcoin treasury in the world. The bet was simple: Bitcoin's long-term appreciation would exceed the cost of debt and dilution.
For a while, it worked. During the 2020-2021 bull run, the strategy produced outsized returns. MSTR shares traded at a premium to net asset value (NAV). Saylor became a celebrity in crypto circles. He appeared on podcasts, gave speeches, and wrote manifestos. His message was consistent: Bitcoin is digital gold, it will compound at 15% annually, and you don't need to worry about it.
But the market cycle turned. By mid-2025, Bitcoin was trading around $75,000, near Strategy's average cost. The leverage began to bite. The company reported a net loss of $8.22 billion in Q2, driven by impairment charges on its Bitcoin holdings. MSTR stock dropped 40% year-to-date, erasing years of gains. And then Saylor sold.
The sale itself was small relative to the total position, but the signal was deafening. A man who repeatedly said he would never sell, who mocked others for doing so, who built a cult of personality around HODLing, was now a seller. The contradiction is not just rhetorical; it is structural.
s heart.
Core: Systematic Teardown of the Strategy
Let me break this down as an auditor, not a fan. I have spent years dissecting leverage strategies in crypto. I wrote a Python script in 2020 to simulate liquidation cascades in Compound Finance. I published a 15-page whitepaper on the fragility of algorithmic interest models. The same cold logic applies here.
1. The leverage is not magic.
Strategy's model works in one direction only: Bitcoin price goes up. When Bitcoin rises, the equity value of MSTR grows faster than the debt, because the debt is fixed in dollar terms. The NAV per share increases. This is positive convexity. But when Bitcoin falls, the same leverage works in reverse. The debt remains, the equity shrinks. The NAV per share drops faster than Bitcoin. Negative convexity. The Q2 loss of $8.22 billion is not a bug; it is a feature of the design.
2. The cost of leverage is hidden but real.
Every time Strategy issues convertible bonds, it pays interest. Every time it sells shares via ATM, it dilutes existing holders. The real cost of the strategy is not the purchase price of Bitcoin, but the spread between Bitcoin's return and the cost of capital. For years, that spread was positive. Now it is narrowing. The company's cash flow from its original software business is minimal; the primary source of cash to service debt is either new equity issuance or Bitcoin sales. The sale we saw was likely a liquidity necessity, not a strategic pivot.
3. The "15% annual return" is a narrative, not a law.
Saylor's claim that Bitcoin will compound at 15% annually is based on historical price performance from 2011 to 2024. But that period included a 1000x from $1 to $70,000. Extrapolating that forward is a fallacy. The law of large numbers applies. As Bitcoin's market cap grows, the percentage gains decrease. A 15% return on a $1 trillion asset is $150 billion in new money each year. That is not impossible, but it is not guaranteed. The narrative is a marketing tool, not a financial model.
4. The single-point failure is real.
Strategy holds 4% of all Bitcoin that will ever exist. That is a massive concentration of risk. If the company is forced to sell a significant portion—say, due to debt covenant breaches or margin calls on any hidden leverage—the market impact would be severe. The recent sale already shows willingness to sell. The trust in the "never sell" narrative is broken. Investors can no longer assume the supply is locked.
5. The custody model is opaque.
Strategy does not self-custody its Bitcoin. It uses third-party custodians. The specific addresses, security arrangements, and insurance coverage are not fully disclosed. For a company claiming to be the ultimate Bitcoin bull, the lack of transparency is a red flag. In my audit experience, centralized custody is the most common failure mode for large holders. The collapse of FTX showed what happens when trust is misplaced. Strategy is not FTX, but the principle applies: code is law until it isn't.
6. The AI advice is orthogonal.
Saylor's advice to young people to learn AI is sound. AI is a transformative technology. But the fact that he gives this advice on the same podcast where he defends his Bitcoin strategy creates a false equivalence. The two are not the same. AI is a general-purpose technology with a clear path to productivity gains. Bitcoin is a volatile asset with an uncertain regulatory future. Saylor is a financier, not an AI researcher. His credibility on AI does not transfer to his Bitcoin strategy.
s heart.
Contrarian: What the Bulls Got Right
I am not here to dismiss Bitcoin entirely. The bulls have points that need acknowledgment.
First, the institutional adoption of Bitcoin is real. The launch of spot ETFs in 2024 brought in billions of dollars from traditional investors. The network effect is stronger than any other crypto asset. Bitcoin's hash rate is at an all-time high, indicating security. The supply cap is enforced by the protocol, not by any human.
Second, Saylor's strategy did work for a long time. From 2020 to 2024, MSTR outperformed Bitcoin itself. The leverage magnified gains. The narrative attracted capital. The company's cost basis was lower than current prices for much of that period. The sell-off in 2025 does not erase the previous success.
Third, the AI advice is genuinely good. Learning to work with AI, understanding its limitations, and building systems that leverage it is a smart career move. Saylor's emphasis on finding the "S-curve" early is valid. AI is at the early stage of an S-curve.
But the contrarian view must be tempered. The bulls ignore the structural fragility of the leverage strategy. They treat Saylor's past success as a guarantee of future returns. They conflate Bitcoin's potential with Strategy's execution. The two are not the same. The disconnect between Saylor's narrative and the company's financial reality is growing.
Takeaway: Accountability Before Narrative
Investors need to track the data, not the words. Watch the quarterly BTC holdings. If Strategy sells two quarters in a row, the strategy has changed. Watch the NAV discount or premium. If MSTR trades below its Bitcoin holdings, the market is pricing in a risk premium. Watch the cash flow. If the company cannot service its debt without selling Bitcoin, the leverage is a liability.
Saylor's AI advice stands on its own. His Bitcoin strategy is a cautionary tale. The narrative of "15% annual return, no worries" is a marketing slogan, not a financial model. The market is finally waking up to that.
s heart.
