OpenAI is reportedly in discussions to lease a massive data center campus in Ohio that could eventually draw as much as 10 gigawatts of power, according to a report from The Information.

The proposed site would mark one of the largest AI infrastructure projects publicly linked to OpenAI so far. The Information reported that Nvidia could be involved in backing the effort, though the details of any partnership were not disclosed. The report indicated that the talks are still ongoing, and no final agreement has been announced.

A project of this scale would underscore how aggressively major AI companies are chasing computing capacity. Training and running advanced AI models requires large clusters of specialized chips, extensive cooling systems and huge amounts of electricity. Data center campuses built for that purpose have become a critical part of the industry’s expansion plans.

If completed, a 10-gigawatt campus would be unusually large by data center standards. The figure points to the potential scale of the power demand involved, though the report did not provide a timeline for construction or operation. It also did not specify the exact location within Ohio or identify the property owner.

The reported talks come as OpenAI continues to expand the infrastructure needed to support its products and future model development. The company has been associated with a broader push across the AI sector to secure more compute, with rivals and partners alike competing for access to chips, energy and land suitable for large-scale facilities.

Nvidia’s name in the report is notable because the company is a central supplier of AI hardware and one of the most influential firms in the sector. However, the extent of any role in the Ohio project remains unclear based on the available information. The report did not say whether Nvidia would be a financial investor, a strategic partner or simply involved in another capacity.

OpenAI has not publicly detailed the project, and the reported lease talks do not guarantee that the campus will move forward. Large infrastructure deals can change as companies evaluate costs, power availability, permitting and construction logistics.

Even so, the discussion reflects the scale of investment now flowing into AI infrastructure. As demand for model training and inference grows, companies are increasingly looking beyond traditional office or cloud setups and toward dedicated campuses capable of supporting industrial levels of power consumption.

For Ohio, a project of this size would be consistent with a wider trend of data center development in markets with available land and access to electricity. Whether the OpenAI talks lead to a deal, however, remains uncertain.

The report did not include comments from OpenAI or Nvidia, and neither company has publicly confirmed the negotiations.