OpenAI reportedly signs massive $300B cloud deal with Oracle

If a company expects to make $12.7 billion in revenue this year, does it make sense to sign a $300 billion purchase order?

In the AI industry, OpenAI is apparently ready to do just that. According to a report from The Wall Street Journal, OpenAI has signed a landmark deal with Oracle to purchase $300 billion worth of cloud computing power over roughly five years, starting in 2027.

It is easily one of the biggest cloud contracts in tech history, and it already reshuffled the global billionaire leaderboard.

Oracle's $317B backlog explosion

The clues for this massive transaction actually dropped earlier this week during Oracle's earnings call.

On Tuesday, Oracle CEO Safra Catz announced that three unnamed customers had signed four multi-billion-dollar deals in the first quarter. Those contracts are part of a surge that Catz said is driving Oracle's cloud infrastructure revenue up by 77 percent this year.

Even more staggering was Oracle's remaining performance obligations. The company added more than $317 billion in future contract revenue in Q1 alone. The news sent Oracle shares into orbit, briefly propelling chairman and co-founder Larry Ellison past everyone else to become the world's richest person.

Oracle declined to comment on the report, while OpenAI did not respond to requests for comment.

Moving past Microsoft to feed the compute monster

OpenAI's partnership with Oracle did not happen overnight. The AI lab first started tapping Oracle for computing power in the summer of 2024.

By January of this year, OpenAI officially broke away from using Microsoft Azure as its exclusive cloud provider. Compute needs have grown so desperate that, according to Reuters, OpenAI even inked a cloud deal with Google this spring, despite competing fiercely against the search giant in frontier AI.

This $300 billion commitment plugs directly into the ambitious Project Stargate. In July, OpenAI and Oracle revealed plans to construct data centers delivering 4.5 gigawatts of power, part of a broader $500 billion domestic data center push alongside SoftBank that was announced with President Trump. Until now, nobody knew how much OpenAI was actually paying to rent that footprint.

At the same time, OpenAI is hedging its hardware bets. The company is reportedly the buyer behind a recent $10 billion contract with Broadcom to design custom AI chips.

Why it matters

We think this contract lays bare the sheer scale and high-stakes financial tightrope of the generative AI race.

For the broader cloud industry, Microsoft's era of exclusive custody over OpenAI is firmly in the rearview mirror. By spreading billions across Oracle, Google, and Azure, OpenAI is protecting itself from single-provider bottlenecks while forcing cloud titans to bid aggressively against one another. Oracle played its cards brilliantly here: by prioritizing raw GPU clusters and bare-metal infrastructure early, it turned what looked like legacy enterprise baggage into the premier AI factory floor.

However, the financial gap is hard to ignore. OpenAI is projected to generate $12.7 billion in revenue this year. Committing to an average of $60 billion annually starting in 2027 assumes astronomical revenue growth that has yet to be fully validated by enterprise adoption. If frontier models do not deliver hundred-billion-dollar enterprise returns by the late 2020s, commitments of this magnitude could quickly morph into a massive overhang for both the startup and its infrastructure hosts.