Oracle co-CEO Clay Magouyrk and OpenAI CEO Sam Altman highlighted rapid software adoption during their keynote at Oracle AI World, but omitted details on how the cloud giant intends to finance its massive $664 billion contract backlog.
Key Takeaways
- Oracle recorded negative free cash flow of $5 billion in its fiscal first quarter as capital spending surged to $28.5 billion, up from $8.5 billion a year earlier.
- OpenAI accounts for roughly half of Oracle's total $664 billion backlog, according to estimates by MarketWise.
- To support operations, Oracle executed a $20 billion common stock sale during the quarter.
- Internally, Oracle has deployed 130,000 active ChatGPT users and over 95,000 Codex users, with AI now writing more than 70% of Oracle's own software.
Software Adoption vs. Capital Expenditure
During the keynote presentation, Sam Altman noted that GPT-6 generated OpenAI's strongest adoption rate to date, citing heavy usage across enterprise software, software engineering, cybersecurity, and scientific research. The integration is already operational within Oracle's corporate environment; a case study revealed that Oracle's talent acquisition team reduced research time by 98% using OpenAI's tools. Furthermore, Oracle plans to host Jalapeño, a custom OpenAI chip designed to execute trained AI models, on Oracle Cloud Infrastructure (OCI).
Despite the product announcements, neither executive addressed the mounting financial commitments behind the expansion. While Magouyrk previously told analysts that Oracle closed over $30 billion in new AI contracts without requiring additional capital, corporate filings highlight significant spending pressures. The $28.5 billion quarterly capital expenditure surge mirrors broader market trends where infrastructure providers face mounting financial burdens, similar to how Meta launched a $35B stock sale amid surging AI capex and Alphabet faced market pressure over capital investments. Oracle reported to CNBC that its recent database enhancements aim to increase software revenue relative to rented hardware capacity.
Why It Matters
Oracle's silence on capital structure exposes a fundamental divide between model creators selling software capabilities and infrastructure providers bearing heavy capital deployment costs. If demand for OpenAI's tools remains robust and translates into high-margin recognized revenue, Oracle can comfortably service its debts and cash outflow. However, with half of Oracle's revenue backlog tied directly to OpenAI, any revenue shortfall or slowdown in model monetization leaves Oracle heavily overextended on raw datacenter capacity.
