Oracle’s $300 Bn OpenAI Gamble Looks Shakier as NVIDIA Rethinks Deal

The likelihood that NVIDIA will invest far less than the mooted $100 billion in OpenAI is raising uncomfortable questions for Oracle.

Chief among them is whether the ChatGPT developer can honour its five-year, $300 billion contract with Oracle signed in September last year, and whether Oracle should recognise the full value of that deal on its own books.

The concerns aren’t unfounded. While bolstering NVIDIA’s relationship with Taiwanese suppliers, chief Jensen Huang last week said that the company’s proposed $100 billion investment in OpenAI was “never a commitment”. WSJ had also earlier reported that the deal had stalled over OpenAI’s business approach and competition concerns.

Later, in an interview with CNBC, he said, “Everything’s on track.”

As of November 30, Oracle reported $523 billion in remaining performance obligations (RPOs), a closely watched disclosure representing contracted revenue yet to be recognised. The figure was around nine times the company’s revenue over the previous four quarters and included the $300 billion tied to OpenAI. Oracle’s shares surged in early September—at one point jumping 36% in a single day—after earnings showed RPOs had more than tripled quarter-on-quarter.

OpenAI’s ties with NVIDIA have resurrected concerns about circular financing in the AI industry. While some of the $100 billion pledge would flow back to NVIDIA through chip purchases, part of it was expected to help OpenAI pay for services such as Oracle’s cloud-computing capacity.

Oracle, meanwhile, dismissed concerns. “The NVIDIA-OpenAI deal has zero impact on our financial relationship with OpenAI. We remain highly confident in OpenAI’s ability to raise funds and meet its commitments,” it posted on X.

Oracle’s Mounting Troubles

Amid capital concerns and onboarding debt financing, the company has faced calls to strengthen its balance sheet to protect its investment-grade rating. In a filing with the Securities Exchange Commission this week, the company laid out its plans to issue up to $20 billion in “equity distribution agreement” this year, as part of a wider $45–50 billion fundraising effort combining equity and debt to expand its cloud-infrastructure business.

The share sale would dilute the holdings of investors who have already seen Oracle’s stock lose roughly half its value since September’s peak. Yet, given the uncertainty surrounding AI mega-deals, bolstering the company’s equity cushion appears prudent.

Jaspreet Bindra, Founder of AI literacy initiative AI&Beyond, explains what went wrong with Oracle. “The AI cloud game is three hyperscalers—Microsoft, Google, and AWS. Oracle tried to muscle in with a large OpenAI deal at very unfavourable terms to itself,” he tells AIM. Bindra adds that the tech major miscalculated and committed to capital intensity closer to cloud hyperscalers than to a traditional enterprise software firm, without a balance sheet or cash-flow profile designed for such risk. 

“The market first gave a huge thumbs up to Oracle, but then saw through the deal, and the stock crashed. I also think Larry Ellison’s billion-dollar guarantees with his son's media ambitions also have a role to play. Oracle over-reached. Even NVIDIA, the most valuable company on Earth, had to give up its cloud ambitions,” he says.

Oracle has said that its latest capital haul will be used to build additional capacity to meet contracted demand from some of its largest cloud customers, including AMD, Meta, NVIDIA, OpenAI, TikTok, and Elon Musk’s xAI. The announcement comes at a time when investors are increasingly questioning whether the wave of AI-linked capital expenditure will generate adequate returns.

Oracle is also building data centres, financed with new debt, on the strength of its OpenAI contract. Building these AI-focused data centres has pushed Oracle’s free cash flow into negative territory, where it is expected to remain until at least 2030, according to Bloomberg. The company is committed to investing in advanced semiconductors and long-term leasing costs tied to new data centre capacity.

The financial strain is now spilling over into workforce and asset decisions. Oracle is reportedly planning job cuts of up to 30,000 employees to help fund its AI infrastructure push, according to a research note by investment bank TD Cowen. The same report said Oracle is also considering selling some non-core activities as US banks pull back from financing its data centre expansion. 

Oracle’s heavy reliance on a small number of large AI customers appears to have ballooned risk, as rising debt, equity dilution, and a falling stock price created a negative feedback loop in market perception. “Layoffs, even if strategic, reinforced the sense of internal strain rather than confidence in execution,” Bindra adds.

OpenAI’s Cash Crunch Worries Oracle

With a large portion of OpenAI’s expected funding now in doubt, questions are mounting over how it will meet an estimated $1.4 trillion in data centre commitments, including those to Oracle. 

“Oracle must cut losses and go back to its core—stick to databases, SaaS—the things they are good at. Do not play with the big boys of the cloud. Oracle must shift from aggressive capacity expansion to disciplined capital allocation. This means phasing AI infrastructure investments tightly against contracted demand, securing stronger prepayment and long-term commitment structures, and clearly articulating unit economics and cash-flow timelines,” suggests Bindra. 

Reducing customer concentration and emphasising Oracle’s strength in regulated, enterprise, and government workloads would stabilise the narrative. “Above all, Oracle needs to make its AI strategy look predictable, cash-generative, and boring—because markets reward certainty more than ambition when leverage is high,” he adds.

Oracle’s greatest risk, however, may be reputational. By asserting that OpenAI’s payments remain probable, management is staking its credibility on an outcome that has yet to be tested.

Meanwhile, OpenAI is seeking to raise another $100 billion in its latest round. SoftBank is in talks to invest up to $30 billion more after lifting its stake to 11% in December with a $22.5 billion investment. OpenAI is also reportedly in discussions with NVIDIA, Amazon, Microsoft, and Middle Eastern sovereign wealth funds. The ChatGPT owner is also targeting an initial public offering later this year.