Oracle says its long-term infrastructure deal with OpenAI is moving forward as planned and directly disputes reports that question the pace and scale of the datacenter operations tied to the agreement. The company responded after recent market volatility and analyst commentary fueled speculation about whether Oracle could deliver the massive computing capacity required for advanced AI workloads.
Earlier reports suggested Oracle had postponed parts of its datacenter program until 2028. Oracle rejected that view and said it set site selection and delivery timelines together with OpenAI, with all projects continuing to meet contractual milestones. According to the company, no sites required for the agreement have slipped, and planning for future expansion continues alongside current builds.
Oracle signed the cloud contract with OpenAI in September, committing to supply roughly 4.5 gigawatts of compute capacity over the coming years. If delivered as planned, the agreement could begin contributing around $30 billion in annual revenue starting in 2027. However, even a modest shift in individual project timelines would not necessarily alter those targets, given how the contracts align with phased capacity delivery.
At the same time, Oracle does not own most of the datacenters it operates. Partners construct the facilities, and once the power, cooling, and connectivity are available, Oracle installs and manages the infrastructure.
The money involved is still quite substantial. Companies must commit significant upfront capital before launching AI datacenter projects, particularly when the deployments rely on specialized computing hardware. Oracle revealed that its capital spending for the fiscal year 2026 is going to be $50 billion, which is a lot higher than the earlier estimates.That update unsettled investors, contributing to a drop in the company’s share price after its latest earnings call.
Analysts have voiced concern about Oracle’s rising debt load, which exceeded $100 billion earlier this year. In response, company executives emphasized that Oracle has multiple financing options and is exploring models that allow customers to deploy their own chips inside Oracle-operated facilities. That approach could reduce balance sheet pressure while preserving long term revenue opportunities.
It seems that Oracle is concentrating solely on reassuring that their AI infrastructure goals are still achievable despite the continued examination by the market and the analysts.
