Oracle Corporation (NYSE:ORCL) reported on September 10 that its revenue rose 30% in the fiscal first quarter, driven by a 121% jump in cloud infrastructure revenue. Cloud infrastructure is the business that rents computing power to AI companies.
The company also said its remaining performance obligations, which are the value of contracts it has signed but not yet delivered, reached $664 billion, more than seven times the revenue it expects this year. On the back of that backlog, Oracle raised its full-year revenue outlook to at least $90 billion.
CFO Hilary Maxson said that if she had to describe the quarter in one word, it would be “acceleration.”
Carol Gauthier/Shutterstock.com
AI Demand Could Turn Cloud Infrastructure Into Oracle’s Biggest Business:
The backlog is the core of the growth story. Oracle said roughly half of the backlog should turn into revenue within three years, which is more than three times what the company expects to collect in the whole of this year.
According to Gartner, worldwide AI spending is expected to reach $2.59 trillion in 2026, more than half of it on infrastructure, and Oracle has positioned itself as a supplier of that capacity.
The newest contracts are also better for Oracle’s balance sheet. It signed more than $30 billion of new AI cloud deals in the quarter, and management said most of those were structured as prepayments or customers bringing their own hardware. That limits how much of its own money Oracle needs to serve them.
Building the Data Centers Is Burning Cash:
Most of the build-out still comes from Oracle’s own balance sheet. The company spent $28.5 billion on data centers and equipment in a single quarter, more than the cash its business generated. Free cash flow stayed negative, as it was for all of last year, and gross margin fell more than five points in a single quarter, to 61%. A day after the results, Oracle expanded its job-cut plan, a sign of how hard the build-out is pressing on cash.
There is also a concentration risk. OpenAI signed a $300 billion contract with Oracle last year that begins in 2027, and that single deal is close to half of the current backlog. OpenAI is still losing money, so Oracle is borrowing to build capacity today against a payment stream that has not started, from a customer whose ability to pay over five years is not guaranteed.
The concerns of a potential AI bubble present a broader risk. If AI investment slows or hyperscalers pull back spending, the demand Oracle is building for could shrink before the data centers are finished.