Strategy — the company formerly known as MicroStrategy, and the most committed corporate holder of bitcoin on earth — reported a second-quarter net loss of $8.22 billion.
The number is almost entirely one line: an $8.32 billion unrealised fair-value writedown on its bitcoin. The software business had a perfectly ordinary quarter, with revenue of $122.4 million, up 6.9%.
But the loss is not the news. This is:
The board has authorised selling bitcoin
Under a BTC Monetization Program approved at the end of June, Strategy may sell up to $5 billion of bitcoin, with $1.25 billion of that earmarked to build US dollar reserves.
For any other company this would be treasury management. For this one it is a doctrinal event. Strategy's entire identity — the reason it trades at a premium to the coins it holds, the reason Michael Saylor became a figure rather than a CFO — was the promise that it would never sell. Bitcoin went in and did not come out.
Two caveats worth stating clearly, because they are being lost in the framing. An authorisation is not a sale; the company has not disclosed how much of the programme it has used, and it may use none. And a $1.25 billion dollar reserve against a position of this size is a liquidity buffer, not an exit.
But the mechanism now exists, approved by the board, and it did not before.
The position is underwater
Strategy holds roughly 843,700 bitcoin — about 4% of all the bitcoin that will ever exist — acquired at an average price near $75,699, for a cost basis around $63.8 billion.
With bitcoin trading below that average, the holding has been marked around $52.6 billion, an unrealised loss of roughly $11.2 billion.
"Unrealised" is doing real work there and deserves the same fairness in both directions: nothing is locked in unless coins are sold, and the accounting swings that produce an $8 billion quarterly loss also produced enormous paper gains on the way up. The company did not become insolvent this quarter. It became, on paper, a buyer who is currently down on the trade.
Saylor's explanation: AI ate the capital
This is the part with implications beyond crypto.
Asked why bitcoin has fallen, Saylor gave five reasons. The one that travelled is what he calls capital suction: more than $1 trillion being drawn into AI data-centre construction.
He named the sinks — SpaceX, Google, Meta, Anthropic, OpenAI, and bitcoin miners themselves pivoting their facilities to AI infrastructure — and argued the pull is visible in both equity and private credit markets. He also cited global trade tensions and tariff uncertainty, and a restrictive Federal Reserve limiting liquidity over the past nine months.
His framing: "This is a capital rotation, not a Bitcoin impairment. Volatility creates opportunity."
Treat that as what it is — the interested view of the person with the largest position — and it is still a claim worth taking seriously, because it is testable and it rhymes with what other balance sheets are showing. The AI buildout is consuming capital at a scale that has to come from somewhere. Miners converting rigs to GPUs is not a narrative; it is a supply-side fact about where the electricity and the capex are going.
Whether that fully explains bitcoin's price is a different question, and one nobody can answer from a single quarter.
What is actually unresolved
Whether Strategy uses the authorisation, and how much. Whether the premium its shares carry over the underlying bitcoin survives a company that can now sell. And whether the capital Saylor says rotated into AI infrastructure rotates back — or stays, because the data centres are still being built.
This article describes what a company reported and what its executives said. It is not investment advice, and nothing here is a recommendation to buy or sell anything.
