Oracle Has a $638 Billion AI Backlog, and Now It Has to Fund It
Oracle has quietly become one of the biggest winners of the AI cloud boom. Its backlog of contracted future business hit 638 billion dollars, up more than 360 percent, powered by mega-deals with OpenAI and Meta. The challenge now is turning that huge promise into cash while spending enormous sums to build the data centers behind it.
Oracle has quietly become one of the biggest winners of the AI cloud boom. Its backlog of contracted future business hit 638 billion dollars, up more than 360 percent, powered by mega-deals with OpenAI and Meta. The challenge now is turning that huge promise into cash while spending enormous sums to build the data centers behind it.
The backlog is the headline. That figure represents revenue Oracle has already signed but not yet delivered, and a jump of this size shows how much AI companies are willing to commit to for cloud computing power. A central piece is Stargate, the AI infrastructure project with OpenAI worth more than 300 billion dollars. Oracle has become a key landlord for the AI boom.
Delivering it means building at industrial scale. Oracle is planning for more than 10 gigawatts of computing power to come online over the next three years, and it is raising 45 to 50 billion dollars this year alone to expand its data centers. It has guided for cloud revenue to grow around 60 percent next quarter. The demand is booked, now the company has to build the capacity to serve it.
That is where the strain shows. All that construction costs cash upfront, and Oracle reported roughly 13 billion dollars of negative free cash flow over the past year, meaning it is spending far more than it takes in as it races to build. The backlog is a promise of future profit, but only if the buildout is funded and the customers pay as expected. Growth this fast is expensive.
The honest question is whether the backlog turns into durable profit. A signed contract is not the same as delivered, funded revenue, and a company burning cash to chase AI demand is betting that demand holds and that it can finance the gap. If AI spending slows or financing gets harder, that enormous backlog becomes a liability as much as an asset. The bet is huge on both sides.
So Oracle has ridden the AI wave to a staggering backlog, and now faces the hard part of paying for it. A 638 billion dollar pipeline, the Stargate deal, 10 gigawatts planned, and heavy negative cash flow. The AI cloud gold rush is real, and so is its cost. Watch whether Oracle can fund the buildout and convert the backlog into cash.