Michael Saylor: The Treasury Company

Say your software company has a good year. There's cash in the account, more than you need to run the place. What do you do with it?

Most operators leave it. It's safe. It's there for a bad quarter. Lessons three and four of the money course showed you what that safety costs: the unit it's counted in loses purchasing power every year, on purpose, by design of whoever holds the pen.

One operator looked at his pile one summer and decided the pile itself was the risk. Then he did something no public company had done at that size.

This is game film, so we're not here for the biography. We're here for the play: what did he do, in what order, what did it cost, and what's the one rule worth taking. The question: what do you do with a pile of cash when you believe the cash is the problem?

Monday, March 20, 2000. MicroStrategy, a software company in Virginia, announces it has to restate two years of revenue. The stock opens at $333 and closes at $193. Its founder, Michael Saylor, has just lost more paper wealth in a day than almost anyone in history. Before the year is out, he and his co-founder settle with the SEC. Fines, disgorgement, no admission.

Hold that scene. It matters for what comes later, because the man who runs the play twenty years on has already watched a fortune disappear once and kept the company.

Tuesday, August 11, 2020. The same company, still run by the same man, still selling the same kind of software. It puts out a press release. It has bought 21,454 bitcoin for $250 million and made it, in the company's own words, the principal holding in its treasury reserve strategy.

The stated reason, from Saylor's statement that day: "This investment reflects our belief that bitcoin, as the world's most widely adopted cryptocurrency, is a dependable store of value and an attractive investment asset with more long-term appreciation potential than holding cash." The company itself was worth a bit more than a billion dollars at the time. It had just put close to a quarter of that into a single asset.

Later he handed the CEO job to his president and made himself executive chairman. His job description became the bitcoin strategy. The play isn't "a company bought bitcoin." The play is "a company turned itself into a machine for buying bitcoin," and the man who built it took himself off everything else.

Ask the tycoon's question. Who made the rule? A founder with a controlling vote, who'd already survived one wipeout. Who had what to gain? Shareholders, if the line went up, and the founder most of all. Who paid? Anyone who lent the company money or bought its paper, if the line went down. Keep that ledger open. We'll settle it on the board.

Strip the play to its parts and it's a loop with five gears. Each one only turns if the one before it did, and the order is the lesson: most imitators copy the later gears without the earlier ones and wonder why the machine grinds.

Gear one is the premise. The company calls its cash a melting ice cube, which means an asset that loses value every year it sits still, because the unit it's counted in is being debased. You learned that mechanism in the money course. The treasury company takes it literally: holding cash is a decision, and it's a losing one.

Gear two is the swap. Cash becomes bitcoin. The software business keeps running, but the balance sheet is now mostly coin. The stock stops being a software stock and becomes a claim on a pile of bitcoin, with a software company attached.

Gear three is the part most people miss, and it's the engine. When investors want bitcoin exposure inside a stock, they'll sometimes pay more for the shares than the bitcoin behind them is worth. That gap is the premium to net asset value, which just means the stock price above the coin per share. When the premium is there, the company sells new shares and buys more coin with the money.

Run the arithmetic. Say each share is backed by one coin and the market pays the price of two coins for it. The company sells a new share, takes the two coins of cash, and buys two coins. Now there are two shares and three coins: a coin and a half behind each one, up from one.

Nobody was tricked; the market paid the premium willingly. Selling stock at a premium to buy the asset is accretive: the holders get diluted in shares and enriched in coin. That's the whole trick, and it only works while the premium lasts.

Gear four is the paper. Once the equity engine is proven, the company borrows against the same idea. First convertible notes, which are loans that can turn into shares.

Then preferred stock, a security that pays a fixed dividend ahead of the common shareholders but never has to be repaid. Strategy issued one series after another. The biggest, STRC, raised over two and a half billion dollars in a single week last summer, sold at ninety dollars against a hundred-dollar face, paying a monthly dividend the company can adjust.

Bitcoin bought with money that costs a fixed yield, held on the belief the coin returns more than the yield. That's a carry trade, run by a corporation, at a scale nobody had tried.

Gear five is the one the play never advertised. Dividends on preferred stock come due every month whether bitcoin went up or not. If the premium closes and the stock can't be sold accretively, the money has to come from somewhere. In August 2026 it came from the coin. A machine built to only buy has, by construction, a day when it has to sell. That day is on the board.

Read the filings, not the podcasts. Strategy publishes an 8-K almost every Monday with the count.

From 21,454 coins in August 2020 to 845,050 on September 7, 2026. Total paid: $63.73 billion. Average cost, fees included: $75,412 a coin. Bitcoin this week trades near $76,700. Six years in, the pile is worth roughly what was paid for it. That's not a verdict; the road in between had the stock up many times over and down by more than half more than once. It's where the score stands today.

Now the week that matters. From August 3 to August 9, 2026, the company sold 1,690 bitcoin at an average of $64,262, and the filing says exactly why: the proceeds were used to buy back STRC preferred stock. It also reports a USD reserve of several billion dollars, described as intended to support the dividends on the preferred stock and the interest on the debt.

Put those two lines together. The company sold coin below its own average cost to retire paper it had sold at a discount to face. The following month it bought back more of the same preferred and didn't buy or sell any bitcoin at all. Gear five, running.

The incentive map, then. Who made the rule? The founder, and the board that approved every series of paper. Who gained? Common shareholders during every stretch the premium was open and the coin was rising, and the founder most, because he held the most.

Who paid? So far, the preferred holders who took ninety cents on the dollar and are now being bought out, and anyone who bought the common at the top of a premium that later closed. Who's still paying? That's what the dividends decide.

Here's the reveal, and it isn't that the play is a fraud, and it isn't that it's genius. It's that both camps are describing the same machine. The treasury company works exactly as long as the market pays a premium for a thing you could buy yourself without the premium.

When the premium is there, every share sold makes the holders richer in coin. When it's gone, the machine has no gear for "wait," only "sell." The founder who lived through that Monday in March built a company that can't sit still.

Game film is for taking one play, not the whole playbook. Here's how the office reads Saylor.

First, the premise stands on its own. You don't need a Nasdaq ticker to ask what your company's cash is a claim on, and whether the unit is melting. Every operator with a surplus faces gear one. The lesson before this one in the school gives you the ruler: measure your money against a unit that can't be issued, every quarter, and see what it did.

Second, the engine isn't available to you, and pretending it's has ruined people. You can't sell shares at a premium to yourself. You can't issue preferred stock. When a retail investor "does what Saylor does," he usually means he borrowed against the coin, which is gear four without gear three, which is just leverage. The office holds no leverage. That rule isn't caution; it's the difference between owning the asset and owing on it.

Third, the rule the desk took. Saylor's yardstick is the thing worth stealing: he measures everything against bitcoin, not against dollars, and so does his company's own scorecard. The desk's paper fund scores itself the same way, against simply holding Bitcoin, so a rising dollar number never gets mistaken for a rising position.

The paper fund owns the stock, sized under its rules, and reads the Monday filing before it reads anyone's opinion of it.

• MicroStrategy was founded in 1989 by Michael Saylor, Sanju Bansal and Thomas Spahr. On March 20, 2000 it announced a restatement of 1998 and 1999 results; the stock fell from $333 to $193 that day, about 42 percent. In December 2000 Saylor and Bansal settled with the SEC, paying $350,000 each in fines and $10 million in disgorgement combined. Wikipedia, Strategy Inc.
• On August 11, 2020 MicroStrategy announced the purchase of 21,454 bitcoin for $250 million and made bitcoin "the principal holding in its treasury reserve strategy." Saylor's quoted statement: "This investment reflects our belief that bitcoin, as the world's most widely adopted cryptocurrency, is a dependable store of value and an attractive investment asset with more long-term appreciation potential than holding cash." The company's market value was a little over $1.2 billion. CoinDesk; Nasdaq press release.
• On September 19, 2022 the company held approximately 130,000 bitcoin acquired for an aggregate $3.98 billion. Wikipedia.
• Saylor became executive chairman on August 8, 2022; Phong Le became CEO. The stock joined the Nasdaq-100 on December 23, 2024. The company was renamed Strategy Inc. on February 5, 2025. Wikipedia.
• The STRC initial public offering priced July 24, 2025 at $90 per share on a $100 stated amount, 28,011,111 shares, and closed July 29, 2025 with gross proceeds of about $2.521 billion and net proceeds of about $2.474 billion; the initial monthly dividend was $0.80 per share. Strategy press releases.
• As of November 17, 2025 the company held over 650,000 bitcoin, then worth roughly $59.69 billion. Wikipedia.
• For the period August 3 to 9, 2026 Strategy sold 1,690 bitcoin at an average $64,262 for $108.6 million; holdings fell to 840,447 bitcoin at an aggregate cost of $63.36 billion and an average $75,385; the filing states the proceeds "were used to fund repurchases of STRC Stock under the Digital Credit Securities Repurchase Program," and it repurchased 1,152,020 STRC shares for $108.6 million. The USD reserve was $4.65 billion, "intended to support the payment of dividends on Strategy's preferred stock and interest on its outstanding indebtedness." Form 8-K filed August 10, 2026.
• For the period August 31 to September 7, 2026 Strategy neither bought nor sold bitcoin, held approximately 845,050 bitcoin at an aggregate $63.73 billion and an average $75,412, repurchased 1,810,885 STRC shares for $176.3 million, raised the preferred repurchase authorization from $1.0 billion to $2.0 billion, and reported a USD reserve of $5.10 billion and USD cash of $1.44 billion. Form 8-K filed September 8, 2026.
• Coinbase BTC-USD spot was $76,724 on September 10, 2026. Coinbase API. Implied: 845,050 bitcoin at that price is about $64.8 billion against $63.73 billion paid.
• Average cost implied at the first purchase: $250 million over 21,454 coins is about $11,653. At September 19, 2022: $3.98 billion over about 130,000 coins is about $30,600. Arithmetic on the sources above.