Strategy Q2 2026 Earnings Call: An $8.2 Billion Loss, a Retired Doctrine, and One Number That Matters More Than Any of It

Strategy reported second-quarter results on 30 July 2026 and held its investor webinar the same evening. The GAAP headline was catastrophic and almost meaningless. The real content was a company explaining, for the first time in full, why it now sells bitcoin — and staking its credibility on getting one preferred share back to $100.

Published 31 July 2026 · Earnings analysis · ~14 min read

1. The 60-second version

Strategy Inc (Nasdaq: MSTR, plus preferreds STRF/STRC/STRK/STRD and STRE on LuxSE) reported Q2 2026 after the US close on Thursday 30 July, then hosted its live webinar at 5:00 p.m. ET. Six things happened:

• A record-scale GAAP loss. Operating loss of $8.33 billion, of which $8.32 billion was an unrealized fair-value markdown on bitcoin. Net loss $8.22 billion, or −$24.45 per diluted share, against net income of $10.02 billion (+$32.60/share) in Q2 2025.
• Holdings still grew. Bitcoin holdings rose 11% during the quarter to roughly 846,000 BTC, and stood at 843,775 BTC as of 26 July — up 25% year to date.
• Debt came down hard. Convertible notes cut from $8.21 billion to $6.71 billion after a May repurchase of $1.50 billion of the 2029 zeros for about $1.38 billion in cash — roughly an 8% discount to par.
• The dividend machine is now expensive. Preferred dividends hit $400.7 million in the quarter, versus $49.1 million a year earlier.
• The never-sell doctrine is formally dead. $218.4 million of bitcoin sold year to date under a board-authorised BTC Monetization Program, mostly to fund preferred dividends.
• Management staked itself to a date. Saylor tied STRC's return to its $100 stated amount to a roughly 70-trading-day repair window pointing at around 8 September.

The stock had closed up 4.73% at $97.74 on Thursday ahead of the release, then eased about 0.45% to $97.30 in after-hours trade — a strikingly muted reaction against options pricing that had implied a move of roughly 8% in either direction.

2. The $8.2 billion loss, and why the number is nearly useless

Every wire headline led with the loss. Almost none of it is cash.

Under ASU 2023-08 (ASC 350-60), effective for US public companies from 1 January 2025, Strategy marks its bitcoin to fair value at each period end and runs the change straight through net income. Bitcoin fell hard through Q2 2026. Therefore Strategy reported an $8.32 billion unrealized loss. There is no judgement, no impairment test and no management discretion in that figure — it is the bitcoin price, multiplied by a coin count, minus last quarter's version of the same sum. The company had already pre-disclosed the figure in a 6 July Form 8-K, which is why the print itself moved nothing.

The more revealing number is the one below the loss: net loss attributable to common stockholders was $8.62 billion, because $400.7 million of preferred dividends sit between net loss and the common line. A year earlier that figure was $49.1 million. The cost of the capital structure has grown roughly eightfold in twelve months, and unlike the bitcoin mark, that one is cash.

3. The bitcoin ledger: 174,895 bought, 3,620 sold

Holdings grew 11% quarter over quarter, from roughly 769,000 to roughly 846,000 BTC, and 25% year to date. Bitcoin Per Share rose 5% in the quarter, from 201,170 to 210,824 satoshis. Year-to-date BTC Yield was 4.5%, BTC Gain 29,997 coins, and BTC $ Gain $1.95 billion at a 27 July reference price of $64,915.

Against 174,895 coins purchased year to date, the company sold 3,620 — a ratio management pointed to as evidence it remains overwhelmingly a net buyer. Both sales were disclosed in detail:

The second sale is worth sitting with. Proceeds of roughly $216 million against a realized loss of $203 million implies the lots disposed of carried a cost basis of around $117,000 per coin — our arithmetic, not a disclosed figure. Strategy did not sell its cheap 2020-vintage coins. It selected its most expensive tax lots, harvesting the loss. Management said the company holds roughly $18.5 billion of unrealized losses for tax purposes, representing a potential $5.4 billion future tax benefit.

That is competent treasury management, and it is also a complete inversion of the doctrine that built MSTR's premium. On the Q1 call Saylor had already signalled the shift, saying a dividend-funding sale would serve "just to inoculate the market". Q2 was the quarter the inoculation became a programme.

Both framings are defensible. The liquidity impact of 3,620 coins on a 843,775-coin position is trivial. The signalling impact of a company that spent four years telling the world it would never sell is not, and management conceded on the call that the sentiment cost of tax-loss harvesting could be negative even where the market impact is negligible.

4. STRC is the whole story now

If you take one thing from this call, take this: Strategy has reorganised itself around defending one preferred security. Roughly half the Q&A concerned STRC, and the CEO, CFO and Executive Chairman all returned to it unprompted.

The Variable Rate Series A Perpetual Stretch Preferred (STRC) is designed to trade near a $100 stated amount. It broke below its trading range on 28 May and was around $89.50 at the time of the call. The response is a five-part defence:

• Dividend rate raised to 12.00%, to be held there until STRC demonstrates sustained trading near $100. Effective yield was described as 13.6%, with a tax-equivalent yield near 21% given return-of-capital treatment.
• USD Reserve grown to $3.75 billion, about 2.1 years of coverage for preferred dividends and interest — inside the stated 2–3 year target, after having been drawn down to roughly half a year of coverage in May.
• A $1 billion buyback programme. Between 20 and 26 July the company bought 288,930 STRC shares ($28.89 million notional) for about $25.00 million, an average of $86.53 — a 13.47% discount to par. Roughly $975 million remains, and management intends to buy more at deeper discounts and taper toward par.
• Return-of-capital tax treatment. The company expects preferred distributions to be treated as non-taxable ROC for the foreseeable future — its language is ten years or more.
• A soft deadline. Saylor noted the dislocation had run 40 trading days and that the team is tracking a roughly 70-day repair window based on the IPO experience, pointing at approximately 8 September.

Why does this matter so much? Because STRC has become the company's primary funding channel. STRC issuance raised $7.53 billion year to date, growth of 254%. Its notional value went from $2.8 billion at the end of Q4 to $10.5 billion at the end of Q2. Institutional holdings rose from $1.1 billion to $3.1 billion, or 22% to 29% of the outstanding, and management said STRC is now the largest single holding in three major preferred indices. Digital credit made up 44% of the $17 billion of capital Strategy raised across the first seven months of 2026.

Saylor framed the ambition in characteristically unhedged terms — the company, he said, is "playing for $1 trillion" — and described the path there as building credit on top of bitcoin rather than extracting value from volatility.

5. The balance sheet actually got stronger

Buried beneath the GAAP wreckage, the quarter's balance-sheet work was the most conservative Strategy has done in years.

The convertible repurchase deserves attention: buying back $1.50 billion of face for $1.38 billion in cash captures a discount and removes a future equitisation decision. The cash-plus-short-term-investments position is roughly flat once you account for the $736.1 million reclassified into short-term instruments; management is parking reserve capital rather than burning it.

ATM issuance ran $8.41 billion during Q2 and a further $1.28 billion between 1 and 26 July, for $17.06 billion year to date. A $1.0 billion MSTR common repurchase programme exists but has not been used — management said it would only buy the common at a sustained discount to net bitcoin per share, a condition it said has not occurred this year. Management also characterised the equity as trading at only about a 5% premium to net asset value, which it read as market scepticism about the value of the digital credit business rather than about the bitcoin.

6. What management ruled out on the call

The Q&A was unusually clarifying, mostly through what was rejected.

• No borrowing against bitcoin. The CEO said the market for it is neither large enough nor well enough priced, flagged counterparty risk, and said the company avoids margin or mark-to-market loans specifically because they hand short-sellers a narrative. Saylor's addition was blunter: they do not need to.
• No volatility selling. Asked twice about covered calls, cash-secured puts and vol derivatives, Saylor rejected the idea — it would strip value from the equity, fragment liquidity, create tax complexity and introduce counterparty risk. Selling credit against a coin, in his framing, is far more lucrative than selling options against it.
• No new instruments. With 11 credit instruments outstanding, the direction of travel is consolidation, not invention, with STRC as the flagship.
• No rush on the converts. The CFO said all options remain open with no prescribed timetable; the priority is STRC. The CEO gave a concrete example: for the 2028 converts with a put date in September 2027, above roughly $183 on MSTR they equitise; below it they can issue at a premium, sell bitcoin, or refinance — and he noted current convertible terms compare favourably to the last five years.
• No return to 98–99% bitcoin. Perhaps the most consequential admission: Saylor said running the balance sheet almost entirely in bitcoin proved counterproductive, and the company will now actively manage the USD/BTC mix countercyclically — leaning to dollars when bitcoin trades at a high premium to its 200-week moving average, and to bitcoin at discounts.

On headwinds, management listed competition from AI capital expenditure, trade tensions, Gulf conflict disruption, restrictive Fed policy and delayed regulatory clarity.

7. The software business, briefly

Revenue of $122.4 million was up 6.9% year over year but came in below consensus, which sat between roughly $125 million and $127 million depending on the provider. Gross profit was $81.6 million at a 66.6% margin, down from 68.8% a year earlier as the mix continues shifting from licence to subscription.

In any other company this would be the article. Here it is a rounding error — the segment's entire quarterly revenue is less than a third of the quarter's preferred dividend bill. It matters mainly as the operating cover that keeps Strategy classified as a software company rather than an investment company.

8. Numbers that need care

Three places where the reporting around this call is loose, and where you should check the 10-Q rather than the wires:

• 846,000 vs 843,775. The CEO's quote cites roughly 846,000 BTC for the quarter end; the press release headline cites 843,775 as of 26 July. The difference is consistent with the early-July sale of 3,588 coins not yet fully replaced — meaning holdings dipped after quarter end. Several outlets have printed the two figures interchangeably. They are different dates.
• Two different "unrealized loss" numbers. The July 26 mark shows a $63.69 billion cost basis against $54.77 billion of market value — a roughly $8.9 billion net position loss. Management separately cited approximately $18.5 billion of unrealized losses for tax purposes. Both can be true: the tax figure aggregates losses at the individual lot level and ignores lots still in profit. They are not comparable, and headlines that treat them as the same number are wrong.
• Coverage arithmetic. A $3.75 billion reserve against about 2.1 years of coverage implies annual dividend-plus-interest obligations near $1.8 billion. That squares with $400.7 million of quarterly preferred dividends plus interest. It also means coverage is a function of the reserve, not of earnings — the operating business does not fund the dividend and nobody on the call suggested it did.

9. What to watch between now and Q3

1. STRC's path to $100, and 8 September. This is now a stated, dated commitment. Watch the weekly buyback disclosures: the pace and average price tell you whether the repair is being bought or is happening on its own.
2. Whether bitcoin purchases resume in size. The company reported no bitcoin purchases in the 20–26 July window. A prolonged pause while ATM issuance continues would change what the ATM is actually funding.
3. The USD/BTC mix in practice. Saylor committed to countercyclical management against the 200-week moving average. That is a testable claim, and the next two quarters will test it.
4. The September 2027 convert put. Not urgent, but the $183 equitisation threshold on MSTR is now public, which makes it a level the market will trade around.
5. Whether monetization stays small. Management framed sales as a fraction of a percent of holdings. The authorisation runs to $1.25 billion. If bitcoin stays below the $75,476 average cost, the gap between "small" and "structural" narrows.

10. FAQ