Strategy Inc

Strategy Inc (Nasdaq: MSTR) is a Delaware-incorporated company listed on Nasdaq. Until 2020 it was known as MicroStrategy — a maker of enterprise analytics and business intelligence software that had operated in that market since the 1990s. It has since adopted a strategy that turned it into an entirely different kind of entity.

The company reports two operating segments. The first is software: product licenses, cloud subscriptions, technical support, and consulting and education services — approximately $122.4 million of quarterly revenue. The second is the bitcoin segment, which generates zero revenue: it sells nothing and produces no income; it holds an asset. As of 7 September 2026 the company held 845,050 bitcoin, acquired at an aggregate cost of $63.73 billion, an average of $75,412 per coin.

Alongside the common stock, five series of preferred stock trade publicly: STRF, STRC, STRK and STRD on Nasdaq, and STRE on the Luxembourg Stock Exchange (euro-denominated). Four carry dividend rates of 10%–12% and one carries 8%. In addition, six series of convertible senior notes are outstanding, with aggregate principal of $6.71 billion.

The company publishes a weekly current report detailing every bitcoin purchase and sale, every share issuance, and every repurchase — alongside a public dashboard on its website. This level of disclosure is unusual, and it makes continuous tracking between quarterly reports possible.

The report shows a net loss of $20.76 billion for the first half of 2026, and $8.22 billion for the second quarter alone. Almost all of it comes from a single line: an unrealized loss on digital assets of $22.77 billion for the half. Since January 2025 the company has applied ASU 2023-08, which requires bitcoin to be measured at market value through the income statement. The consequence: the bottom line moves with the bitcoin price each quarter and does not measure business activity. In the prior-year quarter the same line showed a gain of $14.05 billion.

Setting that line aside produces the picture of the software business on its own: gross profit of $164.9 million for the half against operating expenses of $195.4 million — an operating loss of $30.5 million. The software business does not cover its own operating costs.

Income statement — selected lines ($ millions)

The multi-year revenue series shows a business that contracted over a decade — from $510.8 million in 2021 to $463.5 million in 2024 — and returned to modest growth in 2025 ($477.2 million, +3.0%) and in the first half of 2026 (+9.4%). Within the mix a clear shift is under way: cloud subscriptions grew 54.0% in the quarter, while product support fell 22.7% and licenses fell 48.9%. Per company management, subscription growth is driven by existing customers converting from on-premises installations to the cloud, alongside expanded usage and new contracts.

The balance sheet as of 30 June 2026 shows total assets of $52.56 billion, against $61.64 billion at the end of 2025. The decline stems from the fall in the market value of digital assets, which fell from $58.85 billion to $49.67 billion even as the coin count rose over the period from 672,500 to 846,000.

Two balance sheet movements stand out. First, retained earnings reversed from $6.32 billion to an accumulated deficit of $15.20 billion in six months. Second, the deferred tax liability was almost entirely eliminated, from $1.93 billion to $1.4 million. In the 6 July current report the company explained that the cost basis of its bitcoin exceeded its market value, and accordingly a valuation allowance was recorded against the deferred tax asset, offsetting it in full.

Capital structure — who ranks ahead of whom

This is the section that warrants the most attention in this company, because two layers of senior claims sit between the asset and the common shareholder.

Convertible notes — $6.71 billion

A material fact: the lowest conversion price across all six series is $149.77. The stock traded at $138.96 on 8 September 2026 — that is, every series was out of the money on that date. The report defines the put dates as an unconditional right of holders to require repurchase for cash at 100% of principal. Consistent with that, the report itself presents the debt maturity schedule by put date rather than by stated maturity.

Preferred stock — $15.46 billion liquidation preference at 30 June 2026

Three provisions in the preferred terms worth knowing:

• The liquidation preference per preferred share is the greater of $100 or the trading price (Note 10, footnote 1). STRF trades at $104.20 — meaning its liquidation preference has already risen above $100. The mechanism ratchets upward and does not fall below $100.
• STRD is non-cumulative. If the board does not declare a dividend, it is cancelled and does not accumulate. Its holders also have no right to elect a director upon non-payment.
• Per the report (p. 1954), a deferred dividend may be treated as an increase in the liquidation preference of STRF, STRC, STRE and STRK. Deferral therefore does not eliminate the obligation; it converts it into a larger senior claim.

The USD Reserve and the USD Cash pool

In December 2025 the company established a "USD Reserve" whose purpose, under a board-approved policy adopted in June 2026, is the payment of preferred dividends and interest on indebtedness only. The policy requires a minimum balance equal to 12 months of expected such payments. In August 2026 a separate pool, "USD Cash", was established, which is not restricted to that purpose.

The report notes that the reserve is not subject to any contractual mandate or lien requiring it to be maintained for that purpose — only to board policy. The report further notes (p. 1942) that under Delaware corporate law, dividends may be declared only out of "surplus", measured as assets at market value less liabilities, or out of the year's net profits — and in the company's own words, "we may not have sufficient surplus... even if we have cash and cash equivalents available in our USD Reserve".

A note on cash flow. Operating cash flow for the first half of 2026 was positive, at $9.85 million, following two negative years ($53.0 million in 2024 and $67.2 million in 2025). The report explains the source of the improvement on page 55: "Net cash provided by operating activities increased by $47.2 million... This increase was primarily driven by a $34.4 million increase in interest income (expense), net, mainly attributable to interest earned on the USD Reserve, and a $9.2 million decrease in cash paid for income taxes." Excluding the two items the company identifies, underlying operating cash flow for the half was approximately negative $33.7 million (our computation based on the report's own explanation).

The company reports two reportable segments. The official segment disclosure (pp. 30–31):

This structure is the single most important fact for understanding the company: the segment holding 94.5% of assets produces no revenue at all, and the segment producing all of the revenue is under one percent of the balance sheet. The bitcoin segment's income line is an accounting remeasurement and does not reflect any sale or activity.

Software segment — revenue composition

Per company management, subscription growth reflects existing customers converting from on-premises installations to the cloud service, expanded usage among existing customers, and new contracts. The decline in product support reflects, per the same explanation, that same conversion — a customer moving to the cloud stops paying separate support fees. The two lines describe one substitution, not two separate developments.

The company competes in two entirely separate arenas.

In software it is a long-established participant in a business intelligence market now dominated by far larger cloud vendors. The revenue series tells the story: from $537 million at its historical peak (2011) to $477 million in 2025. The shift to cloud stabilises the base, but scale is not materially expanding.

In the bitcoin treasury arena the company occupies a position that is distinctive in several factual respects: it holds the largest corporate bitcoin treasury in the world; it is included in major equity indices, giving institutional vehicles indirect exposure; and it has issued five separate listed credit series that no competitor has replicated. The report documents the scale of that access: in the single week of 17–23 August 2026, shares were sold for net proceeds of $2.01 billion.

That said, the mechanism underpinning this capability is price-contingent. The report itself explains (p. 55) that one reason its headline metric declined during the quarter was that "our class A common stock [traded] at a lower premium relative to the value of our underlying bitcoin holdings", which reduced the contribution of common stock issuances. In other words: the ability to raise capital is always present, but its contribution to the existing holder depends on the relationship between the share price and the underlying net asset value.

Issuance capacity remaining as of 30 June 2026: common stock $24.26 billion · STRC $17.51 billion · STRD $4.01 billion · STRK $2.10 billion · STRF $1.62 billion. Approximately $49.5 billion in total. The STRC capacity is subject to a self-imposed policy the company has published: not to issue STRC at prices below its $100 stated amount per share. The series traded at $97.92 on 8 September 2026.

The first thing to grasp is that Strategy is not a software company that bought bitcoin, and it is not a bitcoin fund that happens to own software. It is a capital structure. Bitcoin is the asset, the capital markets are the engine, and the software business is a historical remnant that carries the listing and contributes roughly $122 million of quarterly revenue alongside a $66.55 billion asset. Any attempt to analyse the company with software-company tools — margins, revenue growth, multiples — measures an entity that does not exist.

The mechanism that drove the company for five years was simple and clever: if the stock trades above the net asset value behind it, you can issue shares, buy bitcoin with the proceeds, and end the process with more bitcoin behind each share than before. It works. It worked very well when the stock traded at multiples of net asset value. The company even created a dedicated metric to measure it — BTC Yield, the change in bitcoin per assumed diluted share.

The point from which an understanding of the current situation begins is that this mechanism is highly sensitive to the relationship between share price and net asset value, and that relationship has compressed. The smaller the premium, the less each issuance adds for the existing holder. And when part of the proceeds is directed to purposes other than buying bitcoin, the metric can reverse. The report states this plainly: "Because a portion of the proceeds from sales of class A common stock under our ATM was used to pay dividends and interest rather than acquire bitcoin, those sales increased Assumed Diluted Shares Outstanding... without a corresponding increase in our bitcoin holdings."

The second thing to grasp is what sits between the asset and the common shareholder. $6.71 billion of convertible notes and approximately $14.6 billion of preferred stock — together some $21.3 billion of senior claims. They require, per the company's own disclosure, approximately $1.76 billion of cash per year. The report states explicitly that the company does not expect the software business to generate sufficient operating cash flow to satisfy those obligations over the next twelve months.

From this follows the central question: where does that money come from, every year? There are three routes. Issuing common stock — which enlarges the denominator. Selling bitcoin — which shrinks the numerator. Deferring the dividend — which, per the report, may be treated as an increase in the preferred's liquidation preference. All three touch the common shareholder, though in different ways and to different degrees. There is no fourth route that does not, so long as the operating business does not generate the amount itself.

The third thing is that the company changed direction in June 2026, and that is the pivotal development in this report. On 29 June the board approved a new five-part capital framework: a formal USD Reserve policy, a revised STRC dividend rate policy, a preferred stock repurchase program, a common stock repurchase program, and a bitcoin monetization program. A company that builds such a framework is announcing that it has moved from an accumulation phase into a management phase.

The fourth thing is that the framework was actually used, and one can see exactly how — an advantage of the weekly disclosure. Between 29 June and 9 August, 6,916 bitcoin were sold at prices between $59,256 and $64,262, with the proceeds designated per the filings for payment of preferred dividends. Between 24 and 30 August, 4,603 bitcoin were purchased at $80,318. In parallel, of approximately $4.86 billion raised through share issuance over ten weeks, $369.7 million — 7.6% — was designated for bitcoin purchases; the remainder went to the reserve, the cash pool, preferred dividends, and repurchases of STRC.

The fifth thing, and it belongs on the other side of the ledger, is that the same framework also achieved things that are hard to argue with. The USD Reserve grew from 16 months of coverage to roughly 35, nearly three times the policy floor. Preferred stock was repurchased at prices below its stated amount. And the 2029 convertible notes were repurchased for $1.38 billion against $1.50 billion of principal, a difference of $113.9 million. Each of these retires an obligation for less than it is worth at settlement.

The sixth thing is that the credit layer has healed considerably, and this deserves weight. STRC, the largest series, traded at $71.25 at its 52-week low, at $86.52 as the July repurchase price, and at $97.92 on 8 September. STRF trades above its stated amount. If STRC crosses $100, company policy permits it to resume issuing that series — and per the report, $17.51 billion of capacity remains there, capacity that does not dilute the common shareholder at all.

The seventh thing is a distinction between two metrics that are easy to conflate. Bitcoin per share fell over the past ten weeks, from 227,662 satoshis to approximately 200,996. But net asset value per share — which also counts the cash accumulated and the obligation retired — rose over the same period, and by more than the rise in the bitcoin price itself. Both facts are true and they do not contradict: the money raised did not vanish, it simply does not sit in bitcoin. Citing only the first describes an expense that did not occur; citing only the second ignores that an instrument designed to give bitcoin exposure holds less bitcoin per share.

The eighth follows directly: the instrument is changing in character. The ratio of bitcoin exposure to the equity remaining for the common shareholder stood at approximately 1.66 on 30 June and approximately 1.27 on 8 September (our computation from the report and the filings). That is deleveraging. It reduces downside sensitivity and reduces upside participation to a similar degree. Someone who examined this instrument a year ago and someone examining it today are not examining the same instrument.

The ninth thing is the convertible note calendar. The six series carry coupons from zero to 2.25% — very low-cost debt. But all were out of the money on 8 September 2026, and holders have an unconditional right to demand cash repurchase at the put dates, which cluster between September 2027 and September 2028. The report itself presents the maturity schedule on that basis. This is a horizon to mark on the calendar, not an immediate risk.

And the last: there is a circular mechanism here worth holding in mind. The STRC dividend rate is 12% and can be reduced — but per the certificate of designations, by no more than 25 basis points per dividend period and not below monthly SOFR, and only if all accumulated dividends have been paid in full. The report itself warns that successive reductions may depress the price of the series. Company policy, meanwhile, prohibits issuing STRC below $100. Saving on coupon cost and preserving the issuance channel therefore pull in opposite directions. This is a structural tension, not a problem created by error.

Risks as they emerge from the report

• Absolute asset concentration. 90.5% of assets are a single asset. The report notes that the bitcoin price is volatile and that a significant decline in its value could affect the ability to satisfy financial obligations and liquidity needs.
• The company's own statement on cash flow. On page 68: "As of June 30, 2026, we do not expect our enterprise analytics software business to generate sufficient cash flow from operations to satisfy our financial obligations or liquidity needs over the next twelve months." Note carefully: this is a risk factor drafted by the company, and it is not a going-concern qualification from the auditor. KPMG included no such qualification.
• The Delaware surplus test. Preferred dividends may be declared only out of surplus measured at market value, or out of the year's profits. With an accumulated deficit of $15.20 billion, the test rests on the value of the bitcoin. The report notes there may not be sufficient surplus even where cash is available in the reserve.
• Concentration in one series. STRC alone represents 67.8% of the liquidation preference of the entire preferred layer, and carries the highest coupon.
• Dual-class share structure. Class B shares, held principally by the founder, confer control that does not correspond to economic ownership.
• Ongoing dilution. Remaining issuance capacity of approximately $49.5 billion as of 30 June 2026, against a market capitalisation of approximately $58.4 billion (our computation).
• Absence of preferred voting rights. STRC, STRE and STRD confer no right to elect a director even upon non-payment of dividends.
• Regulatory risk. The report details legislation, regulatory initiatives and evolving interpretations concerning digital assets, in the United States and elsewhere.

Items to monitor — what to check in the next filing

Methodological note. The framework below describes conditions, not forecasts. It contains no prices, no probabilities and no targets. Scenarios are descriptive, not predictive. Its purpose is to organise the questions worth tracking — not to reach an investment conclusion.

Conditions that would need to hold for the mechanism to work again in favour of the common stock

• If STRC trades sustainably in the $99–$100 range, then company policy permits resuming issuance in that series, providing a financing channel that does not increase the common share count.
• If a larger portion of issuance proceeds is directed to bitcoin purchases, then bitcoin per share may resume rising. Over the past ten weeks that portion was 7.6%.
• If the common stock repurchase program is activated, then the share count would decrease rather than increase. The program has been authorised since 29 June 2026 and remains entirely unused.
• If the software business turns cash-flow positive excluding interest on the reserve, then part of the carry cost would be funded internally rather than from issuance.

Conditions that would indicate continuing structural pressure

• If bitcoin sales continue under the monetization program, then the source of liquidity for servicing the preferred remains the asset itself.
• If the STRC dividend rate is raised above 12.00%, then the cost of capital in the credit layer is rising. Each additional 50 basis points on the size of that series equals roughly $52 million per year.
• If a dividend on any series is deferred, then per the report the deferral may be treated as an increase in the liquidation preference — the obligation is enlarged rather than saved. On STRD, which is non-cumulative, the dividend is cancelled.
• If months of reserve coverage approach the policy floor of 12 months, then board authorisation is required for any further reduction, and an alternative source would be needed.

Conditions around the put dates

• If the stock trades below the conversion prices (the lowest being $149.77) approaching the put dates, then holders would plausibly prefer cash repurchase over conversion, and the company would need a cash source of the relevant magnitude.
• If the stock trades above the conversion prices, then conversion into shares becomes available — removing the cash obligation while increasing the share count.
• Historical anchor: from 30 Jun 2025 ($107,752) to 30 Jun 2026 ($58,714) the bitcoin price fell 45.5%; and from 30 Jun 2026 to 8 Sep 2026 it rose 34.1%. Looking at those two periods, the quarterly range of movement in the asset exceeds every other unknown in the model.

Six open questions, identical across every company review. They are intended to structure the examination — not to reach a conclusion. This framework is intended to structure analysis, not to produce an investment conclusion.

1. Over ten weeks the common share count rose 13.1% while bitcoin holdings fell 0.1%. From 371.6 million shares and 846,000 coins at 30 June 2026, to approximately 420.4 million shares and 845,050 coins at 7 September 2026. Bitcoin per assumed diluted share fell correspondingly from 210,824 satoshis to approximately 187,742 (our computation from the current reports; the company has not yet reported the third quarter).
2. Of approximately $4.86 billion raised through share issuance in that period, $369.7 million — 7.6% — was designated for bitcoin purchases. Per the footnotes to the current reports, the remainder was designated for the USD Reserve, the USD Cash pool, preferred dividends, and repurchases of STRC stock.
3. The positive operating cash flow for the half ($9.85 million) is attributed in the report primarily to interest earned on the USD Reserve. The report states that the $47.2 million improvement came from $34.4 million of interest and $9.2 million of lower cash taxes. Excluding those two items, underlying operating cash flow for the half was negative (our computation). In 2024 and 2025 annual operating cash flow was negative, at $53.0 million and $67.2 million respectively.
4. All six convertible note series were out of the money on 8 September 2026, and the put dates cluster between September 2027 and September 2028. The lowest conversion price is $149.77 against a share price of $138.96. The report presents the maturity schedule by put date, with a cumulative $5.91 billion falling within a twelve-month window.

This summary is not a recommendation. It is a factual list of what the report presents. The site does not participate in the decision. The decision belongs to the client.