Oracle finds itself at the epicenter of contradictions between technological optimism and financial reality. At the AI World conference, Oracle and OpenAI presented the integration of ChatGPT and Codex services into enterprise products, yet none of the executives disclosed how exactly Oracle plans to finance its contract portfolio (unfulfilled contractual obligations) worth $664 billion.
This is precisely why shareholders continue to assess the risks: nearly half of this portfolio is tied to OpenAI against the backdrop of the company's negative free cash flow.
A bet on AI software sales
Oracle CEO Clay Magouyrk and OpenAI head Sam Altman focused entirely on deploying technological solutions. Altman noted that GPT-6 became the fastest-adopted model in OpenAI's history, pointing to high demand for the neural network in coding, enterprise software, cybersecurity, and scientific research.
Magouyrk stated: more than 70% of Oracle's programs are now written with the help of AI. The OpenAI report reveals specific usage figures — over 130,000 Oracle employees actively work with ChatGPT, and the number of users of the specialized AI coding assistant Codex has exceeded 95,000 people. Oracle's recruiting team has cut the time spent reviewing candidate profiles by 98%.
The company also announced improved practicality of its databases, shifting focus to software sales and reducing dependence on computing power rentals.
Top managers' silence on finances and debt
Not a single figure on capital expenditures, the contract portfolio, or the financing structure was mentioned on stage. Oracle recorded negative free cash flow of $5 billion at the end of the first quarter of the fiscal year. Capital expenditures over the same period rose from $8.5 billion to $28.5 billion, and the IT giant conducted a $20 billion stock offering.
Nevertheless, Magouyrk assured analysts of the business's resilience: Oracle signed new AI contracts worth more than $30 billion without raising external capital. The backlog of unfulfilled obligations reached $664 billion, with about half of them tied to OpenAI.
Oracle will begin offering clients of its OCI cloud infrastructure OpenAI's specialized Jalapeno chip, designed to run ready-made neural networks. High demand from OpenAI will turn these reserves into real revenue, which could smooth over the lack of details. However, if demand falls, the business's vulnerability will become critical.
The situation clearly illustrates the split in the AI market: algorithm developers sell metric growth to investors, while server providers are forced to report to creditors. Oracle finds itself right in the middle, so the company's bet on software is justified only if OpenAI continues to pay.
Cryptalist analysis: The concentration of half of Oracle's contract portfolio in a single counterparty creates a classic concentration risk that the market is not yet fully pricing in. As long as demand for AI infrastructure persists, the monetization model looks sustainable, but any correction in OpenAI's solvency could instantly reprice Oracle's market capitalization and trigger a chain reaction in the AI infrastructure segment.