On December 11 local time, Oracle, the veteran Silicon Valley tech giant, suffered a heavy blow in the capital market. Its stock price plummeted as much as 16% intraday, marking the largest single-day decline since 2001, with its market value evaporating by $102 billion (equivalent to approximately RMB 724.6 billion). By the close of U.S. stock trading, the stock closed at $198.85, still down more than 10%, and its current market value stands at $566.9 billion.
The immediate trigger for the stock collapse was the company's newly released financial results for the second quarter of fiscal year 2026. Data showed that the company's total quarterly revenue reached $16.06 billion, falling short of analysts' expectations of $16.21 billion. Revenue from its cloud business, the core growth driver, hit $8 billion, representing a 34% year-on-year increase but missing the projected target of $8.04 billion. Even though revenue from AI-driven cloud infrastructure soared 68% year-on-year to $4.1 billion, it failed to offset the market's disappointment.
The financial risks exposed in the earnings report have further alarmed investors. Oracle's free cash flow for the quarter stood at -$13.2 billion, far worse than the market expectation of -$5.2 billion, while capital expenditures surged 41% quarter-on-quarter to $12 billion.
Oracle executives disclosed during the earnings call that the company expects to raise its capital expenditure for fiscal year 2026 to $50 billion, an increase of $15 billion from the previous forecast. Morgan Stanley predicts that its adjusted debt may exceed $300 billion by 2028.
The excessive deep reliance on OpenAI has amplified the uncertainty. In September this year, multiple sources reported that OpenAI signed an agreement with Oracle to purchase $300 billion worth of computing power from the latter over a period of approximately five years. This deal ranks among the largest cloud service contracts in history and will require a power capacity of 4.5 gigawatts.
Although Oracle's management did not directly respond to the cooperation agreement with OpenAI, CEO Safra Catz stated that the company has signed important cloud infrastructure service contracts with numerous top enterprises in the artificial intelligence sector, including OpenAI, xAI, Meta, NVIDIA, and AMD, adding that "customer demand far exceeds supply capacity".
However, OpenAI's expenditures far outstrip its revenue-generating capacity, and the market is increasingly concerned about whether the company has the ability to fulfill the hundreds of billions of dollars worth of contracts. This has also exposed Oracle to significant counterparty risk—should OpenAI run into trouble, Oracle's business will be severely impacted.
Under the terms of the agreement between the two parties, OpenAI is required to pay an annual sum of $60 billion, but its current annual revenue is only $10 billion, less than one-fifth of the required payment. Barclays pointed out that 65% of Oracle's new revenue in September came from OpenAI, and the latter's cash flow situation has cast doubt on the contract's fulfillment. Oracle's Credit Default Swap (CDS) has become a proxy indicator for OpenAI's risks.
Barclays has lowered its target price for Oracle from $330 to $310. A month ago, the bank downgraded Oracle's stock rating to "underweight" and warned that its $11 billion cash reserve might be depleted by November 2026, potentially pushing its credit rating to the verge of "junk bond" status.
A report from JPMorgan Chase indicated that to support the expansion of generative AI and large-scale data centers, the global market will need to raise $5–7 trillion in funds between 2026 and 2030, with the funding gap for investment-grade bonds alone reaching as high as $1.5 trillion, a record high in the history of tech infrastructure financing. The institution reminded the industry that the lessons from the large-scale defaults caused by overbuilding of fiber optics around 2000 are still fresh in memory.
|