Oracle (ORCL 0.94%) has defied the odds and turned its cloud infrastructure business into a major growth engine. Even before its most recent quarterly report, the company's cloud business was growing briskly thanks to booming demand for artificial intelligence (AI) infrastructure. In the first quarter of fiscal 2026, Oracle's cloud infrastructure revenue rose 55% year over year to $3.3 billion.
The big shock was a vast expansion of Oracle's remaining performance obligations, or RPO, which measures contracted revenue that hasn't yet been realized. RPO more than quadrupled in the first quarter to $455 billion, and Oracle greatly raised its cloud infrastructure revenue outlook. By fiscal 2030, Oracle now expects the cloud infrastructure business to generate $144 billion in revenue, up by more than a factor of 10 from the current annual run rate.
Image source: Getty Images.
Hitching its wagon to OpenAI
Following the first-quarter report, major news outlets reported that a gigantic deal with OpenAI was behind a big chunk of this new contracted revenue Oracle disclosed. OpenAI has reportedly committed to spending $300 billion over approximately five years on computing power from Oracle, a staggering sum.
One obvious problem with this deal is that OpenAI doesn't have $300 billion. The company has raised around $60 billion in funding in its entire existence. Revenue is reportedly at a $12 billion annual run rate right now, but losses are enormous. News site The Information has reported that OpenAI expects its cumulative cash burn through 2029 to reach $115 billion.
For OpenAI to actually pay for all the computing power that it's agreed to buy from Oracle, the company will need to raise additional funding at a blistering rate for years. On top of that, revenue needs to continue to explode higher and create the need for all that computing capacity.
Competition is one reason to be skeptical that OpenAI can grow revenue as quickly as it needs to. The company's highly anticipated launch of GPT-5 was a flop relative to expectations, and models from Anthropic and others remain highly competitive. OpenAI doesn't appear to have a competitive advantage in the business of creating and selling frontier AI models.
NYSE: ORCL
Key Data Points
What could go wrong?
Oracle needs to build the infrastructure for OpenAI before it can collect revenue. CEO Safra Catz and Chairman Larry Ellison noted in the first-quarter earnings call that Oracle doesn't own data center buildings, only the equipment, and that it can go from spending on infrastructure to collecting revenue from that infrastructure very quickly. In effect, Oracle can only commit to capital spending when it's sure revenue generation is right around the corner.
This reduces some risk for Oracle, but it doesn't solve the main problem. OpenAI is not going to be able to pay for that infrastructure on a continuing basis unless it can keep raising large quantities of capital from investors.
Here's the worst-case scenario for Oracle:
- Oracle builds infrastructure for OpenAI, and OpenAI takes possession and begins paying Oracle.
- This continues to happen until OpenAI faces a crisis. Maybe its revenue projections turn out to be overly optimistic, or the economy enters a recession, or competition eats away at its market share, or the AI boom turns into an AI bust. Funding at the levels necessary to fulfill its obligations becomes impossible.
- OpenAI can no longer pay for, and no longer needs, the vast amount of computing capacity it agreed to buy from Oracle. Oracle is stuck with a lot of expensive AI infrastructure, and it may not be able to find alternative customers if the AI boom has fizzled. We don't know the exact terms of the deal between Oracle and OpenAI, but if the $300 billion doesn't exist, then it doesn't exist.
Oracle's free cash flow is already in negative territory thanks to an increase in capital spending to support its cloud infrastructure business. Capital spending will need to ramp up further in the coming years to support the OpenAI deal, which could require additional debt. If OpenAI fails to follow through on its side of the deal, it's difficult to see a good outcome for Oracle.
If the AI boom continues unabated through 2030, Oracle's audacious $300 billion cloud deal with OpenAI will look like a genius move. If it doesn't, Oracle could be setting itself up for a disaster.
