Oracle just reminded Wall Street that building the AI future is extraordinarily expensive. The enterprise software giant reported fiscal Q2 2026 earnings on December 10, 2025, and investors responded the next morning by heading for the exits. The stock cratered nearly 11%, with intraday losses reaching as steep as 14-15% before a modest recovery.
The culprit was a capital expenditure figure that made analysts do a double-take: $12 billion in a single quarter. That’s roughly three times the approximately $4 billion Oracle spent in the same period the prior year, and about 50% higher than the roughly $8 billion analysts had penciled in. The selloff erased somewhere between $80 billion and $100 billion in market capitalization in a single trading session.
The numbers behind the panic
Oracle’s top line came in at $16.1 billion, narrowly missing the $16.2 billion Wall Street consensus. Oracle’s long-term debt ballooned to $99.6 billion, a 25% jump compared to the same period a year earlier. In English: for every dollar of quarterly revenue Oracle brought in, it was carrying more than six dollars of long-term debt.
The company has been spending aggressively to position itself as a major player in AI cloud infrastructure. A multi-year partnership with OpenAI is among the crown jewels of that strategy, promising massive AI workloads flowing through Oracle’s cloud.
