Strategy Inc. (MSTR): The Software Relic That Swallowed Wall Street's Bitcoin

I. The $50 Billion Leveraged Flywheel

Every Monday morning for most of the past two years, a short filing has landed on the SEC's EDGAR system from a company in Tysons Corner, Virginia. It reads less like a corporate disclosure than a receipt. Last week's version said the company had sold about $22 million of its own stock into the market and used the money to buy 267 bitcoins, paying close to $83,000 apiece.¹ The same filing gave the running total: 848,517 bitcoins, bought for a combined $64 billion, at an average cost of about $75,400 each.¹ That is roughly four of every hundred bitcoins that will ever exist, held by one listed company that still, technically, sells business intelligence software.

The company is Strategy Inc., known until February 2025 as MicroStrategy. Its software business booked about $477 million of revenue in 2025, less than it booked a decade earlier.² Its stock market value on October 21, 2026 was about $45.6 billion, and with debt added the enterprise value was about $50 billion. Against its software sales, that is an EV-to-sales multiple above 130 times, a number that would embarrass even the most fevered AI start-up.² Nobody is paying that multiple for dashboards. They are paying, or until recently were paying, for something stranger: a public company that converted its treasury, then its balance sheet, then its entire identity into a single asset, and then found a way to borrow and issue stock against that asset faster than anyone in corporate history.

The investor's first instinct is to ask what kind of thing this is. On the evidence of the past six years, Strategy behaves less like a software company than like a leveraged closed-end fund that happens to own a software subsidiary. Its revenue has shrunk about 1% a year over ten years.² Its total assets, by contrast, went from about $660 million in 2015 to $61.6 billion at the end of 2025.² Nearly all of that growth came from money raised in capital markets and turned into bitcoin. Financing inflows were about $22 billion in 2024 and about $25 billion in 2025, and almost every dollar went straight out the investing door again.³²

What made that machine work was a premium. For long stretches, investors valued Strategy's shares at well above the market value of the bitcoin behind each share. Whenever that was true, every share sold bought more bitcoin than it diluted, and existing holders saw their slice of the pile grow. When the premium shrinks, the arithmetic turns. Today the stock trades at about $152, almost exactly half its 52-week high, and at roughly 1.2 times book value.² Book value, under the fair-value accounting adopted in 2025, is now mostly bitcoin marked to market. The premium that powered the flywheel has, for practical purposes, gone.

That sets up the four questions this story has to answer. The first is whether the issuance flywheel can keep turning when the stock no longer trades at a premium to its bitcoin. The second is what happens when the convertible bonds come due, the first holder put arriving in September 2027, if bitcoin is still stuck in a downturn.² The third is whether the old software business offers a real floor or is just a public-company shell that costs money to run. The fourth is whether fair-value accounting, taxes and regulators will eventually break the corporate wrapper around the coins.

To answer them, the story has to go back further than bitcoin. It has to start with a young MIT graduate building computer simulations for a chemical company, pass through one of the most spectacular single-day collapses of the dot-com era, sit for a while in a decade of quiet stagnation, and only then arrive at the August afternoon in 2020 when a sleepy software vendor decided its cash was melting. The man at the centre of every one of those chapters is the same, and his habits of mind explain more about Strategy than any balance sheet does.

Subscribe on LinkedIn

↑ Back to Top

II. MIT Systems Dynamics and the DuPont Simulation (1989–1996)

Michael Saylor wanted to fly. He went to MIT on an Air Force ROTC scholarship, aiming at pilot's wings and perhaps an astronaut's seat, until a doctor heard a benign heart murmur and that future closed.⁴ He graduated in 1987 with two degrees, one in aeronautics and astronautics and one in science, technology and society.⁵ The pairing says something about him. One half was engineering; the other was the study of how ideas and institutions change across centuries. His undergraduate thesis, by his own account, was a mathematical model of a Renaissance Italian city-state.⁶ Even at twenty-one he was trying to turn history into equations.

The tool he used was system dynamics, the discipline Jay Forrester had built at MIT to model factories, cities and economies as webs of stocks, flows and feedback loops. Saylor studied it at MIT's Sloan School and carried it into his first jobs, building simulation models for big industrial companies including DuPont, Dow and Exxon.⁶ At DuPont he worked as a venture manager in 1988 and 1989.⁵ The story that has hardened into company legend is that his models of DuPont's markets pointed to trouble ahead, and that the chemical giant, impressed, gave the young modeller a contract worth about $250,000, money he used to start his own firm.⁷ Fortune described his DuPont stint as a brief job as a computer simulation planner before he co-founded the company in 1989, at age twenty-four, with two MIT friends.⁴

The firm, incorporated in November 1989, was MicroStrategy.⁵ Sanju Bansal, an MIT electrical engineering graduate who had been Saylor's friend there, joined in 1990 to run consulting and became chief operating officer in 1993.⁵ For two decades Bansal would be the steady operator beside the showman, the person who made sure the code shipped while Saylor explained to the world what it meant.

What MicroStrategy sold in those early years was a way to ask questions of enormous databases. Large companies had begun to store every transaction they made, but the data sat in relational databases that were hard for executives to query. MicroStrategy's software sat on top of those databases and let a manager slice the numbers by store, by week, by product, by promotion, and see the answer quickly. The technical name for the approach was relational online analytical processing, or ROLAP. The plain-English version is that MicroStrategy built a very fast translator between a business question and a mountain of raw records.

The client that proved the idea was McDonald's. An early contract, reported at about $10 million, paid MicroStrategy to build applications that analysed how well promotions worked across the chain's restaurants.⁴ Think of what that meant for a hamburger company in the early 1990s: for the first time, a marketing executive could see whether a two-for-one deal in Ohio actually sold more food or simply gave away margin. That is the moment business intelligence stopped being a consulting service and became a product category, and MicroStrategy was one of the companies that defined it.

The business grew quickly enough to leave Delaware for the technology corridor of Northern Virginia, near Washington, D.C.⁷ By the late 1990s headquarters sat on Towers Crescent Drive in Vienna, in the Tysons Corner office cluster. The scale-up shows best in headcount. MicroStrategy had 59 employees at the start of 1995 and 907 at the end of 1998.⁸ That is a fifteen-fold increase in four years, fuelled by the client-server software boom, in which big companies spent heavily to put analytical software on every manager's desktop.

The origin matters to today's investor for one reason. Saylor did not learn business by watching markets set prices. He learned it by building models in which, if the inputs and feedback loops were right, the outputs followed. Thirty years later he would describe bitcoin with the same certainty, as an engineering answer to a monetary problem. Whether that habit of mind is a superpower or a blind spot is one of the threads that runs through everything that follows. The first serious test of it came quickly, when MicroStrategy went public and the market started setting a price on his model of the future.

↑ Back to Top

III. The Dot-Com Crucible and the SEC Settlement (1998–2002)

On June 11, 1998, MicroStrategy's Class A shares began trading on Nasdaq.⁸ The company sold 4.44 million shares for $53.3 million, which works out to $12 apiece.⁵ The stock doubled on the first morning, valuing the company at close to a billion dollars and making Saylor's stake worth about $540 million by the close.⁴ It was a strong start. It was nothing compared with what came next.

Over the following eighteen months MicroStrategy became one of the purest expressions of the dot-com mood. Saylor reframed the company as something larger than a business intelligence vendor: a platform that would push information to anyone, anywhere, through wireless devices and the web. Investors bought the vision. By early 2000 the share price had climbed into the hundreds of dollars and Saylor's paper fortune was measured in billions. He lived in a modest Northern Virginia townhouse but courted coverage of his lavish parties, and pledged $100 million to found a free online university that would offer an Ivy League education to anyone with an internet connection.⁹ He was, for a moment, the most celebrated technology founder in the Washington region and one of the most quoted executives in America.

Underneath the vision sat an accounting problem. MicroStrategy's biggest deals were complex, multi-year packages that mixed software licences, services and sometimes partnerships in which the customer also supplied something back to MicroStrategy. The question in each deal was when the revenue had actually been earned. The company had been answering it aggressively, recognising revenue up front on contracts where significant work remained to be done or obligations remained unsettled.

On March 20, 2000, MicroStrategy announced it would restate its results. Revenue for 1999, previously reported at $205.3 million, would come down to somewhere between $150 million and $155 million, and 1998 would fall as well. Its auditor, PricewaterhouseCoopers, said it agreed with the restatement.¹⁰ Shares fell about 62% that day, a drop of $140, to close at $86.75.¹⁰ Saylor's personal paper loss was put at about $6 billion in a single session, more wealth than most American companies of the day were worth.⁹¹¹ The damage did not stop there. After the company also restated 1997, the stock closed at $33 on April 13.¹² Across 1997 to 1999, the SEC later found, revenue had been overstated by about $66 million out of $365 million reported, with about four-fifths of the overstatement in 1999 alone.¹²

The Securities and Exchange Commission sued Saylor, Bansal and former chief financial officer Mark Lynch for financial fraud and settled on the same day in December 2000.¹² Without admitting or denying the allegations, the three agreed to disgorge about $10 million between them, of which Saylor's share was about $8.3 million, and each paid a $350,000 civil penalty.¹² Fortune later noted that once the accounts were restated, an operating gain of about $28 million across 1998 and 1999 became a loss of about $37 million.⁴ The profitable hyper-growth company the market had paid for had never quite existed.

What came next was survival. The internet bust took most of MicroStrategy's wireless ambitions with it, the stock collapsed far below its IPO-era highs, and the company cut deeply. Fortune later counted roughly a third of the workforce lost in the aftermath.¹¹ The online university never materialised on the promised scale. Saylor kept control and kept the company, a non-trivial achievement given how many of his dot-com peers lost both.

The common story about this episode is that Saylor was a reckless promoter who got caught. The record supports a narrower and more useful version. The SEC case was about revenue timing, not invented customers, and the core software kept selling after the restatement. But the episode left two durable marks. The first was on Saylor himself. He had seen, at close range, how a company's reported earnings could evaporate under an accountant's reinterpretation, and how a stock priced on narrative could lose most of its value in a morning. His later disdain for accounting earnings as a measure of value, and his preference for a hard asset he could count, are easier to understand against that March day.

The second mark is on how an investor should read him today. The founder who now runs the world's largest corporate bitcoin treasury is also a founder who once settled fraud charges over how his company described its own performance. That does not make the current disclosures untrustworthy; Strategy's books today are audited by KPMG, which in its 2024 opinion flagged digital-asset accounting as a critical audit matter but found no material weakness in internal controls.³ It does mean the company's own favourite metrics deserve the same scepticism the SEC applied in 2000. When a company invents a new key performance indicator, as Strategy later did with "BTC Yield", the investor's job is to ask what it leaves out.

For the next two decades, though, there would be little drama to scrutinise. MicroStrategy emerged from the bust a smaller, profitable, conservatively run company. It was about to spend the greatest software bull market in history mostly standing still.

↑ Back to Top

IV. The Thirty-Year Stagnation: The Core BI Business (2003–2019)

In September 2014, MicroStrategy's board approved a restructuring that eliminated 777 positions.¹³ Over 2014, headcount fell from 3,158 to 2,470.¹³ Effective the first of that same month, Saylor's base salary went from $875,000 to $1.¹⁴ The gesture was symbolic, since his wealth was in his shares, but the symbol was apt. The company that had helped invent business intelligence was shrinking just as the software industry around it was entering its most prosperous era.

The decade before had not been idle. MicroStrategy kept selling its platform to large enterprises, and it spun ideas out of its engineering base. One of them was Alarm.com, a home-security connectivity business it had incubated; in February 2009 it sold the unit to investors led by ABS Capital Partners for $27.7 million in cash.¹⁵ Alarm.com later went public and became worth far more than that, which makes the sale look, in hindsight, like a company harvesting a seedling rather than growing a tree. Another was Angel.com, a cloud-based call-centre and voice-application business, which it sold to Genesys in March 2013 for about $111 million, booking an after-tax gain of about $57 million.¹⁶ Those sales strengthened the balance sheet. They also showed a pattern: MicroStrategy was good at building adjacent products and less good at turning them into large standalone businesses inside the company.

The company also tried to ride the mobile and social waves. It built consumer-facing apps under names like Alert and Usher, and Facebook became a million-dollar customer.¹⁶¹¹ Technical ambition was never the problem. Converting technical milestones into durable new revenue streams was. None of the experiments changed the trajectory of the core business.

That trajectory is the most important fact about the software company. Over the ten years from 2015 to 2025, revenue went from about $529 million to about $477 million, a decline of about 1% a year.² Revenue in 2014 had been higher still, at about $579 million. Over the same period, the companies that became the faces of modern analytics, Tableau, Qlik, Looker and above all Microsoft's Power BI, grew enormously by offering software that was easier to adopt, cheaper to start with and delivered through the cloud. Microsoft in particular bundled analytics into the productivity suites its customers already paid for, which is a brutal competitor to face when a large part of your revenue depends on high-priced annual maintenance.

Why did MicroStrategy not follow them? The honest answer is that its economics rewarded standing still. A large share of its revenue came from product support, the annual fees enterprises pay to keep using software they bought years ago. Support is extraordinarily profitable and very sticky, because ripping out an analytics system that feeds hundreds of reports across a Fortune 500 company is painful. But that same stickiness made it rational to defend the installed base rather than cannibalise it with cheaper cloud subscriptions. Even in 2024, support still brought in about $244 million, more than half of all revenue, while cloud subscriptions were about $107 million.³ The transition the industry made in the 2010s, MicroStrategy was still working through in the 2020s.

The quality of that installed base is real. No single customer has accounted for 10% or more of revenue in recent years, and the allowance against bad debts was about $2.2 million on roughly $183 million of receivables at the end of 2024, barely more than 1%.³ Receivables have stretched, from about 50 days of revenue in 2014 to about 157 days in 2025, but the explanation is the shape of enterprise contracts, with large annual renewals billed late in the year, rather than customers failing to pay.² These are blue-chip clients who renew slowly and pay eventually. That is the profile of a utility, not a growth business.

By the end of 2019 the picture was settled. MicroStrategy had 2,396 employees and revenue of about $486 million.¹⁷ Operating profit, which had been above $130 million in 2015, had dwindled to roughly zero.² And sitting on the balance sheet was a pile of cash and short-term investments of about $566 million, close to a full year of revenue.¹⁸ For most of the decade that cash had been the company's comfort blanket, the proof that the dot-com near-death would never be repeated.

The common verdict on this period is that MicroStrategy "missed the cloud". The record supports something more precise. It did not miss the cloud so much as decline to pay the price of switching to it, because its maintenance profits made the old model pay better in the short run. The result was a company with a loyal customer base, no growth, shrinking margins and a large hoard of cash earning almost nothing. Any CEO in that position faces the same question: give the cash back to shareholders, buy growth, or do something else entirely. In 2020, with interest rates at zero and the world in lockdown, Saylor chose something else entirely.

↑ Back to Top

V. The August 2020 Pivot: Burning the Cash Boat

On August 11, 2020, with offices across America still empty and the Federal Reserve buying bonds at a pace never seen before, MicroStrategy issued a press release with a headline no listed software company had written before. It had bought 21,454 bitcoins for a total of $250 million, including fees, and adopted bitcoin as its primary treasury reserve asset.¹⁹ Saylor's quote in the release was not hedged. Bitcoin, he said, was "digital gold – harder, stronger, faster, and smarter than any money that has preceded it."¹⁹

The purchase did not arrive out of nowhere. Two weeks earlier, on July 28, alongside its second-quarter results, MicroStrategy had announced a two-pronged capital allocation plan. One prong was a tender offer to buy back up to $250 million of its own Class A shares, the conventional thing an ex-growth company with excess cash does. The other was to invest up to the same amount in "alternative investments", which turned out to mean bitcoin.¹⁹ The company's reasoning was spelled out in the release: pandemic disruption, unprecedented government stimulus and broader uncertainty "may well have a significant depreciating effect on the long-term real value of fiat currencies."¹⁹ Saylor put it more vividly in interviews, describing the cash reserve as a "melting ice cube" as yields collapsed toward zero.²⁰

His argument deserves a fair hearing, because it was not obviously wrong. MicroStrategy had about half a billion dollars of cash that it did not need to run a business with no growth. Treasury bills paid essentially nothing. If inflation ran even a few percent a year, that cash was losing real value every year it sat there, and buying back stock in a shrinking software company was not much more inspiring. Saylor framed the choice as between a certain slow loss and a volatile asset that he believed would hold value against an expanding money supply. Investors could reasonably disagree with the conclusion. The premise, that idle cash on an ex-growth balance sheet is not free, was sound.

What made the move possible was control. MicroStrategy had two classes of common stock. Class A shares, traded publicly, carry one vote each. Class B shares, held overwhelmingly by Saylor, carry ten votes each and convert into Class A one-for-one.³ That structure, a legacy of the founding era, gave Saylor majority voting power in 2020. A board at a widely held company, facing institutional shareholders who had bought a software stock, might have balked at putting half the company's cash into an asset that had fallen more than 80% from its peak only two years earlier. At MicroStrategy, the decision rested with the man who had made it.

The second purchase came quickly. By September 14, 2020, the company had bought another 16,796 bitcoins for $175 million, bringing the total to 38,250 coins at an aggregate cost of $425 million.²¹ In a matter of weeks MicroStrategy had moved most of its cash pile into bitcoin, and it set up the legal and custodial architecture to hold it, eventually including a dedicated subsidiary, MacroStrategy LLC, and institutional custodians.³ The boat had been burned. There was no plausible way back to being a normal software company with a normal treasury.

Wall Street's first reaction was confusion. A business intelligence vendor was suddenly a bet on a currency many institutional investors were not allowed to own. Enterprise customers wondered what it meant for their software vendor's stability. Software analysts, whose models had no line for digital assets, struggled to say what the shares were worth.

Then something happened that changed the plan from a treasury hedge into a business model. The stock began to trade like bitcoin, only more so. Investors who wanted bitcoin exposure but could not hold the coin itself, because of mandates, custody rules or simple caution, found that MicroStrategy shares gave them a version of it through an ordinary brokerage account and a Nasdaq ticker. Demand for the stock rose faster than the value of the bitcoin it held. In other words, the market began to pay a premium for bitcoin wrapped inside MicroStrategy.

Saylor saw the opening immediately. If investors would pay more for the company's shares than the bitcoin behind them was worth, then selling new shares and buying more bitcoin would increase the bitcoin per share for everyone who already owned the stock. If bond investors would lend cheaply against a volatile stock, borrowing to buy bitcoin would do the same. The cash was gone; the capital markets were now the treasury. The results, for a while, vindicated the gamble in the most visible way possible. By February 2024, before the biggest buying sprees had even begun, Bloomberg was reporting that the bitcoin bet was worth about $10 billion, measured against an original cash pile of roughly half a billion dollars.²⁰ A company that had spent a decade unable to grow its revenue had found something the market valued far more than revenue.

That shift, from deploying the company's own cash to raising other people's money, is the true turning point of the story. The first was a one-time bet with the shareholders' savings. The second was a repeatable machine, and like every machine it had an input it could not do without. The next six years would show how much Wall Street was prepared to feed it, and on what terms.

↑ Back to Top

VI. The Financial Engineering Playbook: Convertibles, ATMs, and the 21/21 Plan

On October 30, 2024, MicroStrategy reported third-quarter results that, as usual, showed a software business losing money and a bitcoin treasury growing fast. The headline was something else. The company announced it would raise $42 billion of capital over the next three years, half through equity and half through fixed-income securities, to buy more bitcoin.²² It called this the 21/21 Plan. At the time, the company held about 252,000 bitcoins.²² The plan implied raising about $14 billion a year, roughly thirty times what the software business sold in a year.

To see how a company with a few hundred million dollars of software sales could even propose that, it helps to take the machine apart. It has three engines: equity sold "at the market", convertible bonds, and, later, perpetual preferred stock.

The premium engine

Start with the equity. An at-the-market, or ATM, programme lets a company sell new shares into the stock market a little at a time, through brokers, at whatever the market price happens to be. MicroStrategy used ATMs on a scale never seen before. The logic depended entirely on the premium. Imagine the company's bitcoin is worth $100 per share, and the stock trades at $200. Selling one new share raises $200, which buys $200 of bitcoin. The company now has more bitcoin per share than before, because it issued a claim worth $100 of bitcoin and received $200 of bitcoin in return. Existing shareholders are richer in bitcoin terms, even though they own a slightly smaller percentage of the company.

MicroStrategy built a metric to measure this: BTC Yield, defined as the percentage change in the ratio between the company's bitcoin holdings and its assumed diluted shares outstanding.²² In 2024, BTC Yield was 74.3%.²³ In plain terms, the bitcoin backing each share grew by almost three-quarters in a single year without any change in the bitcoin price. In 2025, it was 22.8%.²⁴ Those are remarkable numbers, and they explain why shareholders tolerated relentless dilution. The scale behind them is hard to picture. In 2025 alone, Strategy raised about $25.3 billion of capital and reported a "BTC Gain" of about 101,900 coins.²⁴ Translated into coins, the accretion from a single year of issuance was more than twice the entire holdings of the next-largest corporate bitcoin treasury.²⁵ When the 2027 convertibles were called, holders swapped them for about 7.4 million new shares, a reminder that much of the debt was always designed to end up as equity.²⁶

They also need a caveat. BTC Yield is not a yield in any normal sense. It produces no cash. It ignores the debt and preferred stock that rank ahead of common shareholders, so a company could in principle raise its BTC Yield by borrowing heavily. And it only measures accretion if the shares are sold above the value of the bitcoin behind them. A metric invented by management to describe management's own performance deserves exactly the scrutiny that the SEC gave MicroStrategy's revenue in 2000.

Selling volatility to hedge funds

The second engine was convertible bonds, and this is where the cleverness lies. A convertible bond pays a low interest rate but lets the holder swap it for shares at a fixed conversion price. The buyers of MicroStrategy's convertibles were mostly hedge funds running a strategy called convertible arbitrage. They buy the bond, short the stock against it, and then trade the hedge back and forth as the stock moves. The more violently the stock moves, the more money they make from rebalancing. MicroStrategy's stock was one of the most volatile large stocks in America; its one-year volatility today is about 79%.² For arbitrage funds that volatility was not a risk to be avoided. It was the product.

So MicroStrategy could borrow almost for free. In November 2024 it sold $3 billion of convertible notes due 2029 with a zero coupon and a conversion price about 55% above the share price.²⁷ In February 2025 it sold another $2 billion at zero coupon, due 2030.²⁸ It was, in effect, selling the volatility of its own stock to the hedge funds that wanted it and taking payment in the form of interest it did not have to pay. Even after years of borrowing, contractual annual interest on the debt at the end of 2025 was only about $36 million.²

The machine also cleaned up behind itself when it could. In September 2024 it redeemed all $500 million of its 6.125% senior secured notes, the one piece of conventional, asset-backed debt in the structure, freeing roughly 69,000 bitcoins that had been pledged as collateral.²⁹ In early 2025 it called its $1.05 billion of zero-coupon notes due 2027, and with the stock far above the conversion price of about $142, holders converted substantially all of them into shares rather than taking cash.²⁶ That is the best case for a convertible issuer: borrow for nothing, watch the stock rise, and see the debt turn into equity.

The ratings agencies saw the other side. In October 2025, S&P Global Ratings assigned Strategy an issuer rating of B-, deep in speculative territory, citing its concentration in bitcoin, the mismatch between a bitcoin balance sheet and U.S.-dollar obligations, weak dollar liquidity and weak risk-adjusted capital, only partly offset by strong market access.³⁰ Put simply, the agency judged the company good at raising money and poor at paying it back from its own operations.

The plans kept getting bigger. In May 2025 management unveiled what it called the 42/42 Plan, doubling the target to $84 billion of equity and fixed income through 2027.³¹ It also raised its 2025 BTC Yield target from 15% to 25%.³²

The third engine arrived in 2025: perpetual preferred stock. These are securities with no maturity date that pay a fixed or variable dividend and rank ahead of common shares, sold mainly to income-seeking investors. Strategy launched four in quick succession. STRK, priced at the end of January 2025, paid 8% and was convertible into common stock.³³ STRF, in March, paid 10% and was cumulative, meaning missed dividends pile up.³⁴ STRD, in June, also paid 10% but was non-cumulative, so missed dividends are simply lost.³⁵ And STRC, nicknamed "Stretch", came in July with a variable monthly dividend that started at 9% and raised about $2.5 billion in its initial sale.³⁶ A euro-denominated version, STRE, followed in November with a 10% coupon.³⁷ By February 2026 the preferred stack had an aggregate notional value of about $8.5 billion.²

Preferreds solved one problem and created another. They let Strategy raise money without diluting common voting power and without a maturity wall. But unlike zero-coupon convertibles, they demand cash. Phong Le, Strategy's chief executive, put the annual dividend bill at about $750 million to $800 million in late 2025, and it has grown as STRC expanded and its rate rose.³⁸ The software business cannot pay that. So the dividends have to come from new capital, from a cash reserve, or eventually from bitcoin.

That leads to the disconfirming arithmetic at the heart of the whole structure. The premium engine runs only one way. If the stock trades at $200 against $100 of bitcoin per share, selling stock is accretive. If it trades at $90 against $100, selling stock to buy bitcoin makes existing holders poorer in bitcoin terms, and selling bitcoin to buy back stock is the accretive move instead. The convertible engine depends on the stock rising past conversion prices. The preferred engine depends on new money to fund old dividends. All three engines draw fuel from the same tank: investor enthusiasm for the common stock. The company spent four years proving how well that design works on the way up. The two years since have begun to test the other direction, and they have also tested whether the company is still run as a software firm at all.

↑ Back to Top

VII. The Identity Split: Phong Le’s Software Engine vs. Saylor’s Evangelism

On August 2, 2022, in the middle of a brutal crypto bear market, MicroStrategy announced that Michael Saylor would step down as chief executive after thirty-three years and become executive chairman, focused on the bitcoin strategy. President Phong Le would become CEO and join the board.³⁹ The same release reported a second-quarter impairment charge on the company's bitcoin of about $918 million, part of nearly $2 billion of cumulative impairments by mid-2022.³⁹ The timing was not subtle. The company's founder was handing the day-to-day running of the business to a finance-minded operator while he devoted himself full time to the asset that was producing those losses.

The split made explicit something that had been true since 2020. Strategy is two companies sharing one ticker. One is a software vendor with about 1,500 employees, most of them outside the United States, selling analytics to large enterprises.³ The other is a capital-markets operation, run from the top, whose job is to raise money and buy bitcoin. Le's public role straddles both: he runs the software business and the capital raising, while Saylor serves as strategist, evangelist and the company's most recognisable voice. Le has also become the company's most candid explainer of its limits. In a podcast interview in late 2025, he described selling bitcoin as a last resort, a step that would make sense mathematically only if the stock fell below the value of its coins and fresh capital dried up.³⁸ It was a rare moment in which the company's leadership spelled out, in public, the conditions under which its central promise could bend.

The software side has been modernised, slowly. Its cloud product, MicroStrategy ONE, and its generative AI features aim to move customers from on-premise licences to subscriptions. The evidence of progress is mixed but real. In 2024, cloud subscription revenue rose to about $107 million from about $81 million, while licence revenue and support both fell.³ Remaining performance obligations, the contracted revenue not yet recognised, stood at about $455 million at the end of 2024, more than half of it due within a year.³ And after years of shrinking, revenue has started to grow again in recent quarters, by about 10% year on year in the third quarter of 2025 and about 7% in the second quarter of 2026.²⁴⁰ That is the first sustained growth in a long time, driven by subscriptions replacing support. It is also small. Even if the software business grew at 10% a year for a decade, it would remain a rounding error next to the bitcoin.

In February 2025 the company acknowledged that reality in its name. MicroStrategy became Strategy, and the new logo folded a stylised bitcoin "B" into the brand, with orange as the primary colour.⁴¹ The software product kept its old name for customers. The parent company stopped pretending its identity was in software.

Governance shifted too, and here the record cuts in an unexpected direction. Every share Strategy sold through its ATM programmes was a Class A share with one vote. Saylor's Class B shares, with ten votes each, stayed constant. So the more stock the company issued, the more Saylor's voting power was diluted. By the end of 2025 he controlled about 40% of the votes.² By the April 2026 record date, it was 37.6%, held through about 19.6 million Class B shares.⁴² Under Nasdaq's definition, a company in which one holder controls more than half the votes is a "controlled company" exempt from some board-independence rules; Strategy no longer fits that description. Saylor remains by far the largest voice, but for the first time since 1989 he could, in principle, be outvoted.

There have been other costs of having a founder who is also a public figure. In June 2024 the District of Columbia's attorney general announced that Saylor and the company had agreed to pay $40 million to settle a lawsuit alleging that Saylor lived in a Georgetown penthouse while claiming residency in Florida or Virginia, avoiding more than $25 million in D.C. income taxes, and that the company had filed false withholding information on his behalf.⁴³ It was described as the largest income-tax fraud recovery in the district's history and the first case under its amended False Claims Act.⁴³ The settlement was immaterial to Strategy's finances. It was not immaterial to the question of how the company handles its founder's affairs, coming a quarter-century after the SEC case. Related-party dealings since then are narrower, mainly the business use of aircraft affiliated with Saylor, governed by an audit committee policy.⁴⁴

That leaves key-person risk, which at Strategy is unusually economic. At most companies, a founder's departure would hit morale and strategy. At Strategy, Saylor's credibility with retail and institutional investors is part of what generated the premium. His daily social-media posts, conference keynotes and television interviews are not marketing in the usual sense; they are a direct input to the price at which the company can sell its stock. If he were to leave, the software business would carry on, but the capital-raising engine would lose its most persuasive salesman. The question of whether a company can run as both a software vendor and a financial vehicle has, in practice, been answered: it runs as a financial vehicle that happens to own software. What the accounting rules made of that hybrid is a story of its own.

↑ Back to Top

VIII. The Accounting Mirror: From Impairment Penalties to Mark-to-Market Chaos

For four years, MicroStrategy's income statement told a story that was, in one specific way, absurd. Under U.S. accounting rules in force until the end of 2024, bitcoin was treated as an indefinite-lived intangible asset. If the price fell below what the company paid, it had to write the asset down and record a loss. If the price rose, it could not write the asset back up. The rules, in effect, let the books see only bad news.

The consequences were vivid. Each time bitcoin dipped, MicroStrategy recorded impairments, and those impairments ran through operating expenses as if the company had suffered a business loss. By the end of 2022, after the crypto crash, shareholders' equity on the balance sheet was negative, at about minus $383 million, even though the company owned billions of dollars of bitcoin at market value.³ By the end of 2024 the company had taken about $4 billion of cumulative impairment losses, and its bitcoin was carried at about $24 billion against a market value of roughly $42 billion.³ Any investor reading the income statement at face value would have concluded the company was failing, when its main asset had in fact roughly doubled in value.

The Financial Accounting Standards Board changed that. Its rule known as ASU 2023-08, issued in December 2023, requires companies to carry crypto assets at fair value and run the changes through net income.⁴⁵ Strategy adopted it from January 1, 2025.² Overnight the balance sheet came to reflect reality, and by the end of 2025 total assets were $61.6 billion and shareholders' equity about $51 billion.²

When the books started telling the truth, loudly

Fair value solved one distortion and created another. Now every move in bitcoin's price flowed straight into quarterly profit. In the second quarter of 2025, as bitcoin rallied, Strategy reported net profit of about $10 billion, or about $34 per diluted share.² Bitcoin then peaked at about $126,000 in early October 2025 and began to slide.⁴⁶ In the fourth quarter of 2025, Strategy lost about $12.4 billion, or about $43 per share.²⁴ In the first quarter of 2026 it lost another $12.5 billion.⁴⁷ In the second quarter, it lost $8.2 billion.⁴⁰ Over the twelve months to June 2026, the cumulative loss came to about $30 billion.

None of those numbers say anything about how the business performed. They say what bitcoin did between two quarter-end dates. The point is clearest in the cash. Over the twelve years from 2014 through 2025, Strategy's reported net profit added up to a loss of about $6.3 billion, while cash from operations added up to a positive $462 million.² More recently, though, operating cash flow has turned negative, at about minus $53 million in 2024 and minus $67 million in 2025.³² That is the number to watch. It says that the software business plus corporate overhead no longer generates cash at all, before a single preferred dividend is paid.

Strategy's own guidance shows how unpredictable the new regime is. In December 2025, management cut its 2025 earnings guidance to a range running from a $5.5 billion loss to a $6.3 billion profit, contingent on where bitcoin ended the year.⁴⁸ That is not guidance in any usual sense. It is a bitcoin price forecast with an accounting translation. By the second quarter of 2026, the company was no longer offering a full-year number at all.⁴⁰

Volatile book profits once threatened a real cash cost. The corporate alternative minimum tax, passed in 2022, imposes a 15% minimum tax on the book income of very large companies. A company that marks bitcoin to market could, in a strong year, report enormous paper profits and owe real tax on gains it had never realised. This was a genuine threat to Strategy's model in 2025. It was largely defused on September 30, 2025, when the U.S. Treasury issued interim guidance allowing companies to exclude unrealised gains and losses on digital assets from the tax's calculation; Strategy says it intends to rely on that guidance and does not expect to owe the tax on unrealised gains.² The risk has not vanished, since interim guidance can be revised, but it has narrowed from an existential tax trap to a policy risk worth monitoring.

The bigger regulatory question has always been what Strategy legally is. A company whose assets are overwhelmingly a single investment looks, on paper, a lot like an investment company, the kind of vehicle the Investment Company Act of 1940 regulates closely and restricts from issuing debt and multiple classes of stock. Strategy argues that it is an operating company with a bitcoin treasury strategy, and it flags the risk of a contrary interpretation in its filings.² An adverse ruling would be devastating to the structure, since the convertibles, preferreds and dual-class shares are exactly the kind of capital stack investment companies are restricted from using. There is no sign of such action today. But it is a risk that grows rather than shrinks as bitcoin becomes a larger share of the balance sheet.

The accounting story, then, is a mirror. Under the old rules it understated the company; under the new ones it shows the truth too vividly, swinging tens of billions a quarter. Neither version helps an investor decide whether Strategy has a durable advantage, or merely got there first. That requires a different set of tools.

↑ Back to Top

IX. The Moat Analysis: 7 Powers and 5 Forces

By March 2026, 148 publicly traded companies held bitcoin on their balance sheets, together owning roughly 1.16 million coins. Strategy alone held about 64% of that total.⁴⁹ The others ranged from crypto miners to medical device makers to a Tokyo-listed company, Metaplanet (メタプラネット), which by April 2026 held about 40,000 coins and ranked third among corporate holders, behind Strategy and Twenty One Capital.²⁵ Almost all of them borrowed Saylor's playbook. Almost none came close to his scale. The question is whether that gap reflects a durable advantage or simply a head start.

Hamilton Helmer's framework of seven powers asks what lets a company earn returns above its cost of capital for a long time. Apply it to the two halves of Strategy separately, because they give different answers.

For the treasury business, the strongest candidate is scale economies in capital markets. Size breeds liquidity, and liquidity breeds cheaper capital. Strategy's stock and options trade in enormous volume, which is exactly what convertible arbitrage funds need to hedge their positions. A $3 billion zero-coupon convertible is possible for Strategy because hedge funds know they can short its stock in size; a smaller copycat issuing at the same terms would struggle to find buyers. Index inclusion reinforced the effect. In December 2024 Strategy joined the Nasdaq-100, bringing passive buying from funds that track it.⁵⁰ That gave the company access to pools of capital that could not buy bitcoin directly.

The case for a related power, a cornered resource in the form of first-mover position, is weaker than it looks. Being first brought Strategy the institutional relationships, the shelf registrations and the brand recognition that copycats lack. But the doors it opened are now open to others. Index providers have also shown they can close doors as easily as open them. In September 2025, S&P passed over Strategy for inclusion in the S&P 500, and the stock fell.⁵¹ In October 2025, index provider MSCI proposed excluding companies whose assets are 50% or more in digital assets; it decided in January 2026 not to proceed for now, but froze increases in share counts for such companies and opened a broader review of "non-operating" companies.⁵² A power that depends on index committees' goodwill is a privilege, not a moat. The market's reactions make the point. The S&P snub knocked about 2.5% off the shares in a day.⁵³ MSCI's January reprieve added about 6% in after-hours trading, in a stock that had already fallen by nearly half during 2025.⁵² When a company's share price moves on whether a committee will classify it as an operating business, the moat is being drawn by someone else.

Counter-positioning, the power an insurgent gains when incumbents cannot copy it without hurting their own businesses, applied briefly in 2020. No traditional company would follow MicroStrategy into bitcoin. By 2026 that had reversed; dozens did.

For the software business, the only plausible power is switching costs. Large customers do not lightly rip out analytics systems wired into hundreds of reports, which is why revenue has been so stable and bad debts so low. But switching costs without pricing power make for a defensive moat, not a compounding one. Ten years of slowly shrinking revenue in a booming market suggest customers stay but do not expand much. The business has no network effects, no process power its cloud rivals lack, and no brand advantage against Microsoft's Power BI.

Porter's five forces sharpen the picture. The force that matters most is the threat of substitutes, and it has become extreme. In January 2024, the SEC approved the listing and trading of ten spot bitcoin exchange-traded products.⁵⁴ Since then, any U.S. investor has been able to buy a fund that holds the coin directly for a small annual fee, without leverage, without dilution and without a founder's key-person risk. Much of the premium Strategy once enjoyed came from being one of the few regulated ways to own bitcoin in a brokerage account. That scarcity is gone.

Buyer power is high on the software side, since enterprises can and do multi-source their analytics tools and renegotiate at renewals. On the treasury side, the "buyers" are investors in Strategy's securities, and they too have grown more powerful as substitutes multiplied: convertible arbitrage funds can now get bitcoin volatility from ETF options and from other treasury companies. The threat of new entrants in treasury strategies is high in principle, though limited in practice by the liquidity advantage described above.

Weighing the evidence, Strategy's real power is narrower than either its fans or critics usually claim. It does have scale-driven access to cheap capital that its imitators cannot match, as its zero-coupon convertibles show. But that power is conditional on two things it does not control: the price of bitcoin and the market's willingness to value its stock above its assets. The strongest test of a moat is whether it survives a downturn. Over the past year, the market has been running that test.

↑ Back to Top

X. Playbook: Business & Investing Lessons

Go back to the balance sheet of December 2019: about $566 million of cash and short-term investments, a software business that had stopped growing, and an operating profit that had dwindled to roughly nothing. Most boards would have read that as a picture of safety. Saylor read it as a slow-motion liquidation. The first lesson of Strategy is that a stagnant company's cash is not neutral. It is a decision being made by default. MicroStrategy's software had shrunk for years, and the cash sitting beside it earned close to nothing in a world of zero rates. Doing nothing was itself a bet, on inflation staying low and on shareholders tolerating a melting pile forever. Founders and boards should take the point without taking the prescription. The right answer for most ex-growth companies is to return the money, not to buy a volatile asset with it. The lesson is narrower and sharper: in a company that has stopped growing, idle cash is a choice, not a cushion.

The second lesson comes from the hedge funds that bought the zero-coupon convertibles. Most companies treat a volatile share price as a cost, something to be managed down with guidance, buybacks and steady messaging. Strategy did the opposite. It recognised that its volatility was an asset it could sell, and that convertible arbitrage funds would pay for it by lending billions of dollars at no interest at all. For investors, the lesson is to look at who is on the other side of a financing and what they are really buying. The buyers of Strategy's bonds were not betting on its creditworthiness; S&P rated it deep in junk territory. They were buying the wildness of its stock. At Strategy, volatility was not the risk to the financing; it was the collateral.

The third lesson is about venues. For a few years, there was a gap between the investors who wanted bitcoin and the investors who were allowed to hold it. Pension funds, mutual funds and many wealth advisers could not buy the coin, but they could buy a Nasdaq-listed stock. Strategy stood in that gap and charged a toll, in the form of the premium its shares commanded over the bitcoin behind them. Then the gap began to close. Spot bitcoin funds arrived, index committees started asking whether digital-asset treasuries belonged in their benchmarks, and well over a hundred other companies began running the same play. Arbitrage profits are real, but they are rent on a market inefficiency, and rent stops when the landlord's monopoly does. Strategy's premium was the price of a door, and doors get copied.

The fourth lesson is the one the past year has taught most painfully. The arithmetic of BTC Yield is symmetrical, but the story told about it was not. Selling stock at twice the value of the bitcoin behind it made every existing shareholder richer in bitcoin terms, and management reported that accretion as a headline number every quarter. Selling stock at less than the value of its bitcoin makes those same shareholders poorer, and the reported yield shrinks accordingly. In 2024 the figure was about 74%; by the middle of 2026, it was 4.5% for the year to date, down from 9.4% only a quarter earlier.²³⁴⁰ The same formula that compounded holders' stake on the way up began to compound against them as the premium faded. A flywheel that only spins forward is not a perpetual-motion machine; it is a ratchet waiting for the market to change direction.

The fifth lesson is about control. The 2020 pivot happened because one man held the votes. In a widely held company, an institutional shareholder base that had bought a software stock would almost certainly have resisted turning the treasury into bitcoin, and the board would have known it. Saylor's ten-vote Class B shares let him decide alone. That cuts both ways. Concentrated control made a bold transformation possible in weeks; it also meant nobody inside the company could stop it if it was wrong. And the same capital-raising machine that the control enabled has slowly diluted that control away, to well under half the votes. The dual-class structure was the launch pad for the bitcoin strategy, and the bitcoin strategy has been quietly dismantling it ever since. Every founder who admires the speed of the 2020 decision should notice that the machine it built is the thing now eroding the founder's grip.

↑ Back to Top

XI. Bear vs. Bull Case Analysis

On October 21, 2026, Strategy's stock closed at about $152. That was roughly half its 52-week high of about $302, and well above its low of about $82.² Bitcoin itself, which had touched about $126,000 a year earlier, had spent the middle of 2026 far lower; in the second quarter, Strategy sold some of its coins at an average of under $60,000.⁴⁶⁴⁰ Over the twelve months to June, the company reported a net loss of about $30 billion. The market value of the whole company now sits well below the market value of its bitcoin, which at the roughly $83,000 the company paid last week comes to about $70 billion.¹

That gap is not, by itself, proof of a bargain or of a disaster, because the common shareholders do not own all the bitcoin. Ahead of them stand the convertible bondholders, with about $6.7 billion of principal after a repurchase in May 2026, and a preferred stack that was already about $8.5 billion in February and has grown since as STRC sold billions more.⁴⁰² On the other side of the ledger, Strategy holds a dollar reserve of about $4.8 billion.¹ A rough reckoning, then, leaves common shareholders with a claim on bitcoin that is somewhat larger than the stock market's valuation of their shares, which means the stock trades near or modestly below the value of the coins that are truly theirs. That is the state of the flywheel tonight: not broken, but no longer self-powering.

The case for the believers

The bull case starts with the asset. If bitcoin becomes, as Saylor argues, a global monetary reserve, then a hoard of nearly 850,000 coins assembled at an average cost of about $75,400 would be close to irreplaceable. No other company comes within a factor of ten of that position; even the third-largest corporate holder owns about a twentieth as much.²⁵ A believer would argue that a company holding about 4% of all bitcoin that will ever exist deserves a premium of its own, the way a gold miner with the largest reserves commands one.

The second plank is that the financing machine has shown it can adapt. When the stock premium shrank, Strategy pivoted to preferred stock, and STRC in particular has become a large, retail-friendly source of capital, raising about $7.5 billion in the first half of 2026 alone.⁴⁰ When the convertibles traded below par, the company bought back $1.5 billion of its 2029 notes for about $1.38 billion, retiring debt at a discount.⁴⁰ And when investors worried about dividends, it built a dollar reserve that management said in February covered about two and a half years of dividends and interest.²⁴ That is not the behaviour of a company stuck on autopilot. It is the behaviour of a management team that keeps finding new levers.

The third plank is the software business. It is small, but it is real, it has started growing again, and it means Strategy is not a shell. It keeps the public-company infrastructure running, and its blue-chip customer base is a modest source of stability that no pure treasury vehicle has.

The case for the sceptics

The bear case begins with the same facts, viewed from the other side. First, the premium is gone, and the evidence of what that means is already in the numbers. BTC Yield fell from 9.4% at the end of the first quarter to 4.5% by July, which can only happen if new issuance diluted bitcoin per share during the second quarter.⁴⁷⁴⁰ Selling stock to buy bitcoin is no longer clearly accretive. Selling bitcoin to buy back stock or preferreds is starting to make more sense, and the company has begun doing a version of exactly that.

Second, the company that said it would never sell has started selling. In the first half of 2026, Strategy sold about $218 million of bitcoin, all of it to fund preferred dividends, under a programme that also authorises sales to top up the dollar reserve and to buy back securities.⁴⁰ Le's own last-resort test from late 2025 had two conditions, and by the measure that matters most, the first has now been met. The amounts so far are tiny relative to the hoard. But the symbolism matters for a company whose premium rested partly on the promise that its coins were a one-way street.

Third, the cash bill keeps rising while the operating business consumes cash rather than producing it. Contractual interest on the convertibles is small, but preferred dividends run to hundreds of millions of dollars a year, and the STRC rate has climbed to 12%.⁴⁰ The software business, meanwhile, ran an operating cash outflow of about $67 million in 2025.² Every dollar of dividends therefore comes from new capital, from the reserve, or from bitcoin. A reserve that covers a couple of years buys time; it does not change the equation. It is also worth remembering how the reserve was built. When Strategy created it in December 2025, at about $1.44 billion, it funded it by selling common stock through its ATM programme, at the same moment it cut its full-year earnings guidance.⁴⁸ A company whose founding thesis was that cash is a melting asset sold its own shares to hold cash. That was a prudent decision. It was also an admission that the machine needed a dollar cushion to keep its promises to preferred holders.

Fourth, there is the bond calendar. The convertibles do not all mature at once, but they carry dates on which holders can demand their money back. The first is September 15, 2027, on $1 billion of notes due 2028. The next cluster arrives through 2028: $2 billion of 2030 notes in March, the remaining 2029 notes in June, and two smaller issues in September.² In round numbers, close to $6 billion of principal becomes puttable within about a year, starting eleven months from now. Most of that debt converts into stock only at prices far above today's; the 2029 notes, for instance, convert at about $672 a share and the 2030 notes at about $433.²⁷²⁸ If the stock stays near current levels, holders will want cash, not shares. The dollar reserve could cover most of the 2027-2028 puts on paper, but then it would not be available for dividends. The realistic paths are refinancing with new convertibles, issuing more stock or preferreds at whatever price the market offers, or selling bitcoin. None of them is catastrophic at today's prices. All of them get uglier if bitcoin falls further.

An activist looking at this structure would ask pointed questions. Why does a company whose stock trades at or below its asset value keep issuing stock? Why should the software business, which shareholders cannot value separately, sit inside a vehicle whose fortunes it cannot affect? And why does management's preferred performance metric ignore the senior claims that have grown faster than the common equity over the past year? None of those questions has a crisp answer from the company yet.

Three measures matter more than any others for following the story from here. The first is BTC Yield, the change in bitcoin per diluted share. It was about 74% in 2024, about 23% in 2025 and 4.5% for 2026 through July, and the direction is clearly down.²³²⁴⁴⁰ If it turns negative for a full quarter, the core promise of the strategy will have reversed. The second is the market's valuation of the company against its bitcoin, net of the debt and preferred claims ahead of common shareholders. Today it sits near or modestly below parity, down from premiums that once ran to roughly double. Above one, issuance helps shareholders; below one, it hurts them. The third is the cash the business burns before dividends, measured by operating cash flow, which went from slightly positive in 2023 to about minus $67 million in 2025.² Set against the dollar reserve, that number shows how long Strategy can keep paying its preferred holders without selling bitcoin or raising money on poor terms.

Weighed together, the evidence narrows both cases. The bull case does not need bitcoin to go to the moon; it needs bitcoin to rise enough, by 2027 and 2028, to bring the convertibles back into the money and restore a premium. The bear case does not need a crash; it needs a long, grinding sideways market in which preferred dividends and bond puts slowly force the company to sell what it promised to keep. Which of those happens is not a question about Strategy at all. It is a question about bitcoin, with Strategy as the amplifier.

↑ Back to Top

XII. Epilogue: The Horizon of Corporate Alchemy

Tonight Strategy is a company worth about $45 billion that owns about $70 billion of bitcoin, owes billions to bondholders and preferred investors, and runs a software business that most of its shareholders never think about. It sits in a strange middle ground. It is legitimate enough to be in the Nasdaq-100 and to carry a credit rating from S&P. It is unconventional enough that index committees debate whether it is an operating company at all. Its founder still posts about bitcoin every day, still holds the largest block of votes, and still insists that the company will never need to sell its coins in any serious way. The coming months will test every part of that position.

The first test arrives next week. On October 28, 2026, Strategy's shareholders will vote on a proposal to move STRC, its largest preferred stock, from monthly to daily dividends.⁵⁵ It is a small technical change with a large message: the company is trying to make its preferred stock feel as much like a cash account as possible, to keep retail money flowing in. If STRC keeps growing, Strategy has found a funding source that does not depend on the common stock's premium. If it stalls, the company's options narrow quickly.

The second test is the bond calendar. The first holder put falls on September 15, 2027, and the heavier ones follow through 2028. Between now and then, Strategy must decide whether to refinance, repay from its reserve, issue equity at whatever price the market offers, or sell bitcoin. Each choice will reveal something. Refinancing at low rates would show that the volatility trade still works. Repaying from the reserve would show discipline at the cost of dividend cover. Selling bitcoin in size to meet a put would answer the central question the story has carried since 2020: whether a company can borrow against bitcoin forever without ever having to hand any back.

The third test is regulatory. As bitcoin approaches the entire value of the balance sheet, the question of whether Strategy is really an investment company will not go away, and MSCI's broader review of non-operating companies remains open.⁵² A favourable outcome would cement Strategy's status as an ordinary operating company with an unusual treasury. An unfavourable one could force index funds to sell and call the whole capital structure into question.

The fourth is the one that will matter most for the story's legacy. Well over a hundred public companies have copied the playbook.⁴⁹ If bitcoin recovers and they prosper, Strategy will look like the founder of a new corporate form. If bitcoin stagnates and they struggle to pay their own debts and dividends, Strategy will look like a singular event, a once-in-a-generation alignment of a stubborn founder, zero interest rates and a market inefficiency that has now closed.

What remains is the tension that has run through every chapter. Strategy may be the most original idea in corporate finance in a generation, a way to turn an ordinary listed company into a leveraged, liquid, regulated wrapper for a new monetary asset. Or it may be the most elegant leveraged bet ever built, a structure that works brilliantly while prices rise and has never been tested through a long decline. The record so far supports both readings. The next two years will decide which one history keeps.

↑ Back to Top

XIII. Outro

In 1989, in Wilmington, Delaware, a twenty-four-year-old who had wanted to fly sat over simulation printouts for DuPont, convinced that the right model could tell a company what was coming. He built a business on that conviction, watched it nearly die when the market decided its numbers were wrong, and then spent twenty years running it carefully, as if determined never to be caught out again.

Thirty-seven years later, the model is no longer predicting chemical demand. It has turned a public software company into something that has no real precedent: a machine that takes Wall Street's appetite for volatility and converts it, week after week, into a scarce digital asset. Every Monday, the receipt arrives on EDGAR. A few hundred more coins, a few million more shares, a slightly larger pile. Whatever happens to bitcoin, there will never be another company quite like it: a business software vendor that stopped waiting for customers and decided to manufacture its own destiny out of digital scarcity and Wall Street volatility.