On the same evening that Amazon booked a $53.4 billion gain from revaluing a holding that does not trade, MicroStrategy booked an $8.32 billion loss from revaluing an asset that does.
The exact same accounting mechanism, two opposite directions, the same night.
Coinbase: Revenue Fell - and the Structure Improved
| Metric | Q2 2026 | Change |
|---|---|---|
| Revenue | $1.22 billion | -14% sequential, -19% year over year |
| Transaction revenue | $599 million | -21% sequential |
| Subscriptions and services | $632 million | 48% of net revenue |
| Adjusted EBITDA | $208 million | - |
| Net loss | $(359.5) million | - |
| Cash and cash equivalents | $8.6 billion | - |
The Number That Answers the Question We Posed Yesterday
Robinhood reported yesterday that crypto plunged 38%, while the rest of its platform was at a record. And the question we posed was: is this weakness specific to Robinhood, or an industry-wide phenomenon?
The answer arrived: Coinbase's transaction revenue fell 21% in the quarter, and total revenue fell 19% year over year. This is industry-wide.
And to Coinbase's credit, the company notes that the decline in transaction revenue was better than the decline in trading volumes across the crypto market as a whole - meaning it lost less than the market did.
And the structural point: subscriptions overtook trading
This is the first time it has happened: $632 million from subscriptions and services, against $599 million from trading.
And why that matters more than the revenue decline:
Transaction revenue is cyclical by definition. It depends on appetite, on volatility and on mood - which is precisely why Coinbase has always traded as a leveraged bet on the price of bitcoin.
Subscription and stablecoin revenue is something else entirely: it comes from balances sitting on the platform, from interest on them, and from services consumed even when nobody is trading. It does not vanish in a quiet quarter.
And this quarter, for the first time, the stable part is larger than the cyclical part. In the first quarter it was 44% of net revenue; now 48%, and in absolute terms - the larger of the two.
That is exactly what a company needs to happen when its market cools.
MicroStrategy: The Loss That Has Nothing to Do With the Business
| Metric | Q2 2026 | Year ago |
|---|---|---|
| Revenue (software business) | $122.4 million | $114.5 million (+6.9%) |
| Gross profit | $81.6 million (66.6% margin) | $78.7 million |
| Operating income/loss | $(8.33) billion | +$14.03 billion |
| Of which: digital asset revaluation | $(8.32) billion | +$14.05 billion |
| Net income/loss | $(8.22) billion | +$10.02 billion |
| Per diluted share | $(24.45) | +$32.60 |
Consensus stood at a profit of $52.04. The result: a loss of $24.45.
But that 'miss' says nothing about the company
MicroStrategy's software business brought in $122.4 million in the quarter, and grew 6.9%. It is a small business, stable and gross-margin profitable - and it is everything the company actually does.
All the rest - $8.32 billion of loss - is a change in the value of the bitcoin it holds. It did not sell a coin. Nothing happened in the business. The price of an asset fell, and an accounting rule requires that to be recorded in the income statement.
Which is why the comparison to consensus is meaningless in this case. An analyst who forecast $52.04 was not forecasting the business - they were forecasting the price of bitcoin. Anyone reading this stock is reading a bitcoin chart with a layer of debt on top, not a financial statement.
And the Numbers That Do Matter at MicroStrategy
| Metric | Value |
|---|---|
| Bitcoin holding | 843,775 coins - the largest institutional holder in the world |
| Growth in the holding year to date | +25% |
| Capital raised year to date (ATM) | $17.06 billion |
| STRC raised year to date | $7.53 billion (+254%) |
| Cumulative preferred dividends | $1.06 billion |
| Dollar reserve | $3.75 billion - covering more than 2.1 years of dividends and interest |
| Bitcoin sales year to date | $218.4 million |
Two figures deserve particular attention:
The first - the dollar reserve. $3.75 billion covering more than two years of dividend and interest payments. A company funded with debt and preferred stock has to pay cash even when its asset is falling - and this reserve is precisely the defense against the scenario in which it is forced to sell bitcoin in order to pay.
And the second - it has started selling. A "bitcoin monetization program" with $218.4 million of sales year to date. That is a small amount relative to the holding, but it is a change in principle at a company whose brand was "we buy and we never sell."
And the Connection Nobody Else Will Make Tonight
On that exact same evening:
| Company | The item | The amount | Effect on the headline |
|---|---|---|---|
| Amazon | Revaluation of its Anthropic holding | +$53.4 billion | EPS of $5.75 instead of ~$1.95 |
| MicroStrategy | Revaluation of its bitcoin holding | -$8.32 billion | A $24.45 loss instead of ~a small profit |
The exact same accounting rule - mark an asset to fair value and recognize the change in the income statement - produced on the same night the largest gain in Amazon's history and the largest loss in MicroStrategy's.
And this is not an anecdote. It is a widening phenomenon: the more public companies hold large non-operating assets - stakes in private companies, digital currencies, securities - the less the net income line describes the business and the more it describes asset prices.
The practical conclusion: net income has stopped being a good measure of the business across a growing share of the market. Anyone who wants to know what happened at a company needs to read operating income - and the footnotes.
The Bull Thesis
At Coinbase, whoever reads it positively will point to subscriptions overtaking trading - 48% of net revenue, and $8.6 billion of cash. The company is becoming less cyclical precisely as its market cools, and that is the right thing at the right time.
At MicroStrategy, whoever reads it positively will see 843,775 bitcoin and $17.06 billion raised year to date without strain, with a $3.75 billion reserve covering more than two years of payments. For anyone who wants leveraged exposure to bitcoin - this is the instrument.
The Bear Thesis
At Coinbase, revenue fell 19% year over year and the company lost $359.5 million. 48% subscriptions is good, but 52% still depends on appetite, and appetite is cooling.
At MicroStrategy, the structure is the risk: debt and preferred stock demanding ongoing cash payments, against an asset that generates no cash flow. As long as bitcoin rises, that is an engine. When it falls, the reserve is what stands between the company and a forced sale - and that is exactly why it built one.
The debate in one line
The bulls see at Coinbase a successful shift to a subscription business larger than trading precisely as the market cools, and at MicroStrategy the largest institutional bitcoin holding in the world with a cash reserve covering two years. The bears see revenue down 19% and a loss at Coinbase, and a capital structure at MicroStrategy demanding ongoing cash from an asset that generates none. Both sides are reading the same reports.
My Angle
A personal opinion of Ilan Abramov - not advice, not a recommendation
What catches me in these two reports is that they demonstrate two completely opposite things you can do with the same underlying asset.
Coinbase built a business that neutralizes the volatility. Subscriptions, stablecoins, custody services - things consumed even in a boring quarter. And it worked: revenue fell 19%, but the company still generates positive adjusted EBITDA and holds $8.6 billion in cash. In a quarter where trading volume collapsed, that is exactly the difference between a business and a bet.
And MicroStrategy built the precise opposite: maximum leverage on the same asset. That is a legitimate and openly declared choice - but it has to be read correctly. The $8.22 billion loss this quarter tells you nothing about the company, exactly as the $10 billion profit a year ago told you nothing. The only number that describes the business is $122.4 million of software revenue.
And what I find most interesting is actually the $3.75 billion dollar reserve. A company that built a brand on "we buy and we never sell" has assembled a cash cushion covering two years of payments, and started an orderly sales program. That does not contradict the thesis - it acknowledges that a leveraged capital structure demands cash, and that a falling market does not wait for you to recover.
And the point I hold above all the others: on the same night, the same accounting rule gave Amazon $53.4 billion and took $8.32 billion from MicroStrategy. At both of them, net income does not describe the business.
And what I will watch: the subscription share at Coinbase. 44% last quarter, 48% today. If it keeps climbing even once trading recovers - this is a different company from the one we knew. If it falls back the moment appetite returns, then it was not a structural change, it was a denominator that shrank.
Summary
Coinbase reported revenue of $1.22 billion - down 14% sequentially and 19% year over year - with transaction revenue plunging 21% in the quarter to $599 million, adjusted EBITDA of $208 million and a net loss of $359.5 million. And for the first time, subscriptions and services ($632 million) exceed transaction revenue, making up 48% of net revenue.
MicroStrategy reported a net loss of $8.22 billion, or $24.45 per share - against a $10.02 billion profit a year ago. It stems entirely from an $8.32 billion unrealized loss on the bitcoin holding. The software business itself brought in $122.4 million (+6.9%). The company holds 843,775 bitcoin, has raised $17.06 billion year to date, and has built a dollar reserve of $3.75 billion.
And the question we posed yesterday has its answer: Robinhood's 38% crypto decline was not specific to Robinhood. Coinbase's transaction revenue fell 21% in the quarter - this is industry-wide.
Sources: the official results releases of Coinbase Global and of Strategy Inc (formerly MicroStrategy) for the second quarter of 2026, as filed with the U.S. Securities and Exchange Commission on Form 8-K on July 30, 2026, including revenue and its breakdown, operating and net income and loss, the revaluation of digital assets, the bitcoin holding, the capital raises, the dollar reserve and the monetization program; Robinhood figures from its results release of July 29, 2026; Amazon figures from its results release of July 30, 2026; analyst consensus ahead of the reports. Data accurate as of the time of writing. The charts are shown in real time via TradingView. Nothing herein constitutes a forecast, recommendation or advice - see the full disclaimer at the bottom of the page.