Oracle, OpenAI focus on adoption, skip funding at event

At Oracle AI World, both companies showcased ChatGPT and Codex integrated into Oracle's software. The CEOs did not address how Oracle finances its $664 billion contract backlog.

The presentation by Clay Magouyrk and Sam Altman stayed on adoption. Meanwhile, investors are considering that roughly half of the backlog is linked to OpenAI, against a backdrop of negative free cash flow.

Is Oracle Selling AI Software or Renting AI Capacity?

During the keynote, Altman noted that GPT-6 achieved the strongest adoption OpenAI has seen for a new model. He mentioned applications in coding, enterprise software, security, and science.

According to Magouyrk, AI currently generates over 70% of Oracle's own software code.

Separate from that, a case study from OpenAI reports 130,000 active ChatGPT users within Oracle, along with more than 95,000 users of its coding agent Codex.

The same study indicates that Oracle's talent acquisition group reduced its research time by 98%.

On Tuesday, Oracle stated that the new features are intended to make databases more actionable, as reported by CNBC. This aim could move revenue towards software and away from capacity rental.

Why Did the Oracle and OpenAI Pitch Skip Oracle’s Financing?

No figures regarding capital spending, backlog, or financing were mentioned during the on-stage discussion.

Oracle posted a negative free cash flow of $5 billion for its fiscal first quarter, according to its report. CNBC noted that capital spending reached $28.5 billion in that period, compared to $8.5 billion a year earlier. Additionally, the firm executed a $20 billion common stock sale within the quarter.

By contrast, Magouyrk informed analysts that Oracle secured over $30 billion in new AI contracts without requiring extra capital.

The backlog—contracted revenue that has yet to be recognized—stood at $664 billion. MarketWise estimates that OpenAI accounts for roughly half of that amount.

Additionally, Oracle plans to offer Jalapeño, an OpenAI-designed chip for running trained AI models, on its Oracle Cloud Infrastructure (OCI).

Should OpenAI's demand persist, the backlog will turn into revenue, making the financing omission less significant. However, if demand declines, that concentration will become the main focus.

This divide mirrors a broader split in the AI market. Model creators promote adoption, whereas infrastructure providers must address lender concerns. Oracle operates in both areas, so optimism on the software side depends on OpenAI continuing to pay for capacity.